Data management for ITSTGCN

STGCN

Author

SEOYEON CHOI

Published

July 5, 2023

순환형 구조를 가진 모델(Models with recurrent structures):

GConvGRU: 그래프 순환 유닛 (Graph Convolutional Gated Recurrent Unit)
GConvLSTM: 그래프 순환 LSTM (Graph Convolutional Long Short-Term Memory)
GCLSTM: 그래프 순환 LSTM (Graph Convolutional LSTM)
DCRNN: 이중 그래프 순환 신경망 (Dual Graph Convolutional Recurrent Neural Network)

그래프 합성곱 모델(Graph convolution models):

LRGCN: 로컬-전역 그래프 합성곱 신경망 (Local-Global Graph Convolutional Network)
TGCN: 전이 그래프 합성곱 신경망 (Transition Graph Convolutional Network)
EvolveGCNO: 진화형 그래프 합성곱 네트워크 (Evolutionary Graph Convolutional Network with Optimization)
EvolveGCNH: 진화형 그래프 합성곱 네트워크 (Evolutionary Graph Convolutional Network with Hypernetwork)

동적 값 업데이트를 활용한 모델(Models utilizing dynamic value updates):

DyGrEncoder: 동적 그래프 인코더 (Dynamic Graph Encoder)

논문에서 제안하는 방법은 GConvGRU (Chebyshev Graph Convolutional Gated Recurrent Unit Cell)와 GConvLSTM (Graph Convolutional Recurrent Network)이라는 두 가지 모델을 소개합니다. GConvGRU는 그래프 데이터에 대한 시간적인 의존성을 캡처하는데 사용되며, GConvLSTM은 그래프와 시퀀스 데이터를 동시에 처리하는데 활용됩니다.

GConvGRU는 Chebyshev 그래프 합성곱과 Gated Recurrent Unit (GRU)을 결합하여 그래프 데이터의 시간적 의존성을 모델링합니다. 이 모델은 그래프 내 노드들 간의 연결과 그래프의 구조를 고려하여 시계열 데이터를 예측하고 분석하는데 유용합니다.

“GC-LSTM: 그래프 합성곱 임베딩 LSTM을 이용한 동적 링크 예측”이라는 논문은 동적인 그래프에서 링크 예측 작업에 그래프 합성곱 네트워크(GCN)와 Long Short-Term Memory(LSTM) 셀을 결합한 새로운 모델을 제안합니다.

논문에서는 그래프 구조가 시간에 따라 변화하는 동적인 그래프에서 노드들 사이의 링크 존재 여부를 예측하는 문제를 다룹니다. 기존의 링크 예측 방법들은 동적인 그래프의 변화를 처리하는데 어려움이 있습니다.

제안하는 GC-LSTM 모델은 그래프 합성곱을 사용하여 노드의 특징을 임베딩하고 그래프 구조를 파악합니다. 동시에 LSTM 셀을 활용하여 시간적 의존성과 순차적 패턴을 모델링합니다.

GC-LSTM 모델은 동적인 그래프 데이터를 시간 순서대로 처리하여 변화하는 그래프 구조를 적절히 반영하고 미래의 링크를 정확하게 예측할 수 있습니다. 실험과 평가를 통해 GC-LSTM 모델이 동적인 링크 예측 작업에서 다른 최신 기법들보다 우수한 성능을 보여주며, 동적인 그래프 구조를 다루는데 유용한 접근 방법임을 입증합니다.

GNAR stationary condition

import itstgcn
import torch
import pandas as pd

import numpy as np
import random

data_dict = itstgcn.load_data('./data/fivenodes.pkl')
loader = itstgcn.DatasetLoader(data_dict)

dataset = loader.get_dataset(lags=2)

sum_result = 0
for i in range(5):
    sum_result += np.sum(np.array(dataset.targets)[:,i])
print(sum_result)

-28.048902156588262

from torch_geometric_temporal.dataset import ChickenpoxDatasetLoader
loader1 = ChickenpoxDatasetLoader()

dataset1 = loader1.get_dataset(lags=4)

sum_result = 0
for i in range(20):
    sum_result += np.sum(np.array(dataset1.targets)[:,i])
print(sum_result)

9.553510148236436

from torch_geometric_temporal.dataset import PedalMeDatasetLoader
loader2 = PedalMeDatasetLoader()

dataset2 = loader2.get_dataset(lags=4)

sum_result = 0
for i in range(15):
    sum_result += np.sum(np.array(dataset2.targets)[:,i])
print(sum_result)

-39.06545396690812

from torch_geometric_temporal.dataset import WikiMathsDatasetLoader
loader3 = WikiMathsDatasetLoader()

dataset3 = loader3.get_dataset(lags=8)

sum_result = 0
for i in range(1068):
    sum_result += np.sum(np.array(dataset3.targets)[:,i])
print(sum_result)

-1543.6160620151331

loader6 = itstgcn.load_data('./data/Windmillsmall.pkl')

dataset6 = loader6.get_dataset(lags=8)

sum_result = 0
for i in range(11):
    sum_result += np.sum(np.array(dataset6.targets)[:,i])
print(sum_result)

0.33864788949161806

from torch_geometric_temporal.dataset import MontevideoBusDatasetLoader
loader10 = MontevideoBusDatasetLoader()

dataset10 = loader10.get_dataset(lags=8)

sum_result = 0
for i in range(675):
    sum_result += np.sum(np.array(dataset10.targets)[:,i])
print(sum_result)

1114.5722422676997

Import

import pandas as pd
import matplotlib.pyplot as plt

GConvGRU

df_GConvGRU = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GConvGRU.csv')

pedal_wiki_GSO_GConvGRU = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GConvGRU_pedal_wiki_GSO.csv')

df_GConvGRU['model']='GConvGRU'
pedal_wiki_GSO_GConvGRU['model']='GConvGRU'

df_GConvGRU['dataset'].unique()

array(['fivenodes', 'chickenpox', 'pedalme', 'wikimath', 'windmillsmall',
       'monte'], dtype=object)

GConvLSTM

df_GConvLSTM = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GConvLSTM.csv')

pedal_wiki_GSO_GConvLSTM = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GConvLSTM_pedal_wiki_GSO.csv')

df_GConvLSTM['model']='GConvLSTM'
pedal_wiki_GSO_GConvLSTM['model']='GConvLSTM'

df_GConvLSTM['dataset'].unique()

array(['fivenodes', 'chickenpox', 'pedalme', 'wikimath', 'windmillsmall',
       'monte'], dtype=object)

GCLSTM

df_GCLSTM = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GCLSTM.csv')

pedal_wiki_GSO_GCLSTM = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GCLSTM_pedal_wiki_GSO.csv')

df_GCLSTM['model']='GCLSTM'
pedal_wiki_GSO_GCLSTM['model']='GCLSTM'

df_GCLSTM['dataset'].unique()

array(['fivenodes', 'chickenpox', 'pedalme', 'wikimath', 'windmillsmall',
       'monte'], dtype=object)

DCRNN

df_DCRNN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_DCRNN.csv')

pedal_wiki_GSO_DCRNN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_DCRNN_pedal_wiki_GSO.csv')

df_DCRNN['model']='DCRNN'
pedal_wiki_GSO_DCRNN['model']='DCRNN'

LRGCN

df_LRGCN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_LRGCN.csv')

pedal_wiki_GSO_LRGCN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_LRGCN_pedal_wiki_GSO.csv')

df_LRGCN['model']='LRGCN'
pedal_wiki_GSO_LRGCN['model']='LRGCN'

TGCN

df_TGCN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_TGCN.csv')

pedal_wiki_GSO_TGCN = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_TGCN_pedal_wiki_GSO.csv')

df_TGCN['model']='TGCN'
pedal_wiki_GSO_TGCN['model']='TGCN'

EvolveGCNO

df_EvolveGCNO = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_EvolveGCNO.csv')

pedal_wiki_GSO_EvolveGCNO = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_EvolveGCNO_pedal_wiki_GSO.csv')

df_EvolveGCNO['model']='EvolveGCNO'
pedal_wiki_GSO_EvolveGCNO['model']='EvolveGCNO'

DYGRENCODER

df_DYGRENCODER = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_DYGRENCODER.csv')

pedal_wiki_GSO_DYGRENCODER = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_DYGRENCODER_pedal_wiki_GSO.csv')

df_DYGRENCODER['model']='DyGrEncoder'
pedal_wiki_GSO_DYGRENCODER['model']='DyGrEncoder'

EvolveGCNH

df_EvolveGCNH = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_EvolveGCNH.csv')

pedal_wiki_GSO_EvolveGCNH = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_EvolveGCNH_pedal_wiki_GSO.csv')

df_EvolveGCNH['model']='EvolveGCNH'
pedal_wiki_GSO_EvolveGCNH['model']='EvolveGCNH'

GNAR

df_GNAR = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GNAR.csv')

wiki_GSO_GNAR = pd.read_csv('./simulation_results/Real_simulation_reshape/Final_Simulation_GNAR_wiki_GSO.csv')

df_GNAR['model']='GNAR'
wiki_GSO_GNAR['model']='GNAR'

df = pd.concat([df_GConvGRU, df_GConvLSTM, df_GCLSTM, df_DCRNN, df_LRGCN, df_TGCN, df_EvolveGCNO, df_DYGRENCODER, df_EvolveGCNH,df_GNAR]).reset_index().iloc[:,1:]

df

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
0	fivenodes	STGCN	0.0	NaN	2	12.0	NaN	50.0	0.729374	80.985221	GConvGRU
1	fivenodes	STGCN	0.0	NaN	2	12.0	NaN	50.0	0.729082	80.891788	GConvGRU
2	fivenodes	STGCN	0.7	rand	2	12.0	linear	50.0	1.892262	81.976547	GConvGRU
3	fivenodes	STGCN	0.7	rand	2	12.0	nearest	50.0	2.211288	87.803869	GConvGRU
4	fivenodes	STGCN	0.8	rand	2	12.0	linear	50.0	2.072818	103.648742	GConvGRU
...	...	...	...	...	...	...	...	...	...	...	...
26786	pedalme	GNAR	0.8	rand	4	NaN	nearest	NaN	1.302679	0.023105	GNAR
26787	pedalme	GNAR	0.5	rand	4	NaN	linear	NaN	1.302679	0.022839	GNAR
26788	pedalme	GNAR	0.5	rand	4	NaN	nearest	NaN	1.302679	0.023041	GNAR
26789	pedalme	GNAR	0.8	rand	4	NaN	linear	NaN	1.302679	0.060872	GNAR
26790	pedalme	GNAR	0.8	rand	4	NaN	nearest	NaN	1.302679	0.023526	GNAR

26791 rows × 11 columns

df.to_csv('./df_fig.csv')

df2 = pd.concat([pedal_wiki_GSO_GConvGRU,pedal_wiki_GSO_GConvLSTM,pedal_wiki_GSO_GCLSTM,pedal_wiki_GSO_DCRNN,pedal_wiki_GSO_LRGCN,pedal_wiki_GSO_TGCN,pedal_wiki_GSO_EvolveGCNO,pedal_wiki_GSO_DYGRENCODER,pedal_wiki_GSO_EvolveGCNH,wiki_GSO_GNAR])

df2

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
0	pedalme	STGCN	0.300000	rand	4	12.0	linear	50.0	1.557424	6.553443	GConvGRU
1	pedalme	STGCN	0.300000	rand	4	12.0	nearest	50.0	1.591089	7.995071	GConvGRU
2	pedalme	STGCN	0.600000	rand	4	12.0	linear	50.0	1.737521	7.022613	GConvGRU
3	pedalme	STGCN	0.600000	rand	4	12.0	nearest	50.0	1.675156	6.745583	GConvGRU
4	pedalme	IT-STGCN	0.300000	rand	4	12.0	linear	50.0	1.123114	12.278102	GConvGRU
...	...	...	...	...	...	...	...	...	...	...	...
418	wikimath	STGCN	0.511945	block	8	12.0	linear	50.0	0.830234	255.170347	EvolveGCNH
419	wikimath	IT-STGCN	0.511945	block	8	12.0	linear	50.0	0.746624	630.021935	EvolveGCNH
0	wikimath	GNAR	0.511945	block	8	NaN	nearest	NaN	1.353637	143.252073	GNAR
1	wikimath	GNAR	0.511945	block	8	NaN	nearest	NaN	1.353637	211.333467	GNAR
2	wikimath	GNAR	0.511945	block	8	NaN	nearest	NaN	1.353637	218.516156	GNAR

3783 rows × 11 columns

fivenode

Baseline

df.query("model=='GNAR' and dataset=='fivenodes'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26500	fivenodes	GNAR	0.0	NaN	2	NaN	NaN	NaN	1.40683	0.021981	GNAR
26501	fivenodes	GNAR	0.0	NaN	2	NaN	NaN	NaN	1.40683	0.017151	GNAR
26502	fivenodes	GNAR	0.7	rand	2	NaN	linear	NaN	1.40683	0.084960	GNAR
26503	fivenodes	GNAR	0.7	rand	2	NaN	nearest	NaN	1.40683	0.010853	GNAR
26504	fivenodes	GNAR	0.8	rand	2	NaN	linear	NaN	1.40683	0.012061	GNAR
...	...	...	...	...	...	...	...	...	...	...	...
26768	fivenodes	GNAR	0.3	rand	2	NaN	nearest	NaN	1.40683	0.008497	GNAR
26769	fivenodes	GNAR	0.5	rand	2	NaN	linear	NaN	1.40683	0.010377	GNAR
26770	fivenodes	GNAR	0.5	rand	2	NaN	nearest	NaN	1.40683	0.018586	GNAR
26771	fivenodes	GNAR	0.6	rand	2	NaN	linear	NaN	1.40683	0.007493	GNAR
26772	fivenodes	GNAR	0.6	rand	2	NaN	nearest	NaN	1.40683	0.008042	GNAR

204 rows × 11 columns

pd.merge(df.query("dataset=='fivenodes' and mtype!='rand' and mtype!='block'").groupby(['model','nof_filters','lags','epoch'])['mse'].mean().reset_index(),
         df.query("dataset=='fivenodes' and mtype!='rand' and mtype!='block'").groupby(['model','nof_filters','lags','epoch'])['mse'].std().reset_index(),
         on=['model','nof_filters','lags','epoch']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	nof_filters	lags	epoch	mean	std
0	DCRNN	2.0	2	50.0	1.229	0.041
1	DyGrEncoder	12.0	2	50.0	1.114	0.037
2	EvolveGCNH	12.0	2	50.0	1.175	0.068
3	EvolveGCNO	12.0	2	50.0	1.168	0.065
4	GCLSTM	4.0	2	50.0	1.209	0.023
5	GConvGRU	12.0	2	50.0	0.732	0.005
6	GConvLSTM	12.0	2	50.0	1.131	0.041
7	LRGCN	4.0	2	50.0	1.212	0.024
8	TGCN	12.0	2	50.0	1.085	0.016

Random

https://matplotlib.org/stable/gallery/statistics/boxplot.html#sphx-glr-gallery-statistics-boxplot-py

# with plt.style.context('cyberpunk'):
#     plt.rcParams['figure.figsize'] = [40,20]
fig, ax = plt.subplots(3, 3,figsize=(40,20))

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==2 and epoch==50 and model=='GConvGRU' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,0],grid=False,widths=0.5)
ax[0,0].set_title('GConvGRU')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==2 and epoch==50 and model=='GConvLSTM' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,1],grid=False,widths=0.5)
ax[0,1].set_title('GConvLSTM')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==4 and lags==2 and epoch==50 and model=='GCLSTM' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,2],grid=False,widths=0.5)
ax[0,2].set_title('GCLSTM')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==4 and lags==2 and epoch==50 and model=='LRGCN' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,0],grid=False,widths=0.5)
ax[1,0].set_title('LRGCN')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==2 and epoch==50 and model=='DyGrEncoder' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,1],grid=False,widths=0.5)
ax[1,1].set_title('DyGrEncoder')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and lags==2 and epoch==50 and model=='EvolveGCNH' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,2],grid=False,widths=0.5)
ax[1,2].set_title('EvolveGCNH')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and lags==2 and epoch==50 and model=='EvolveGCNO' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,0],grid=False,widths=0.5)
ax[2,0].set_title('EvolveGCNO')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==2 and epoch==50 and model=='TGCN' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,1],grid=False,widths=0.5)
ax[2,1].set_title('TGCN')

df.query("dataset=='fivenodes' and mtype=='rand' and inter_method == 'linear' and nof_filters==2 and lags==2 and epoch==50 and model=='DCRNN' and mrate in [0.3  , 0.5  , 0.6 ,0.7  , 0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,2],grid=False,widths=0.5)
ax[2,2].set_title('DCRNN')


for ax in ax.flat:
    ax.set_yticklabels([])
    ax.set_yscale('log')
    ax.axvline(x=2.5, color='black', linestyle='-')
    ax.axvline(x=4.5, color='black', linestyle='-')
    ax.axvline(x=6.5, color='black', linestyle='-')
    ax.axvline(x=8.5, color='black', linestyle='-')
    ax.set_xticklabels(['IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN'])
    ax.set_xlabel('')
    ax.set_ylabel('')
    
fig.suptitle('',fontsize=40)

Text(0.5, 0.98, '')

pd.merge(df.query("dataset=='fivenodes' and mtype=='rand'").groupby(['model','mrate','nof_filters','inter_method','method','lags','epoch'])['mse'].mean().reset_index(),
         df.query("dataset=='fivenodes' and mtype=='rand'").groupby(['model','mrate','nof_filters','inter_method','method','lags','epoch'])['mse'].std().reset_index(),
         on=['model','inter_method','method','nof_filters','mrate','lags','epoch']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate==0.7 and inter_method=='linear'")

	model	mrate	nof_filters	inter_method	method	lags	epoch	mean	std
12	DCRNN	0.7	2.0	linear	IT-STGCN	2	50.0	1.247	0.044
13	DCRNN	0.7	2.0	linear	STGCN	2	50.0	1.271	0.066
32	DyGrEncoder	0.7	12.0	linear	IT-STGCN	2	50.0	1.252	0.060
33	DyGrEncoder	0.7	12.0	linear	STGCN	2	50.0	1.548	0.158
52	EvolveGCNH	0.7	12.0	linear	IT-STGCN	2	50.0	1.188	0.049
53	EvolveGCNH	0.7	12.0	linear	STGCN	2	50.0	1.228	0.064
72	EvolveGCNO	0.7	12.0	linear	IT-STGCN	2	50.0	1.162	0.052
73	EvolveGCNO	0.7	12.0	linear	STGCN	2	50.0	1.198	0.045
92	GCLSTM	0.7	4.0	linear	IT-STGCN	2	50.0	1.228	0.034
93	GCLSTM	0.7	4.0	linear	STGCN	2	50.0	1.245	0.033
112	GConvGRU	0.7	12.0	linear	IT-STGCN	2	50.0	1.180	0.060
113	GConvGRU	0.7	12.0	linear	STGCN	2	50.0	1.858	0.139
132	GConvLSTM	0.7	12.0	linear	IT-STGCN	2	50.0	1.287	0.075
133	GConvLSTM	0.7	12.0	linear	STGCN	2	50.0	1.472	0.125
152	LRGCN	0.7	4.0	linear	IT-STGCN	2	50.0	1.244	0.041
153	LRGCN	0.7	4.0	linear	STGCN	2	50.0	1.261	0.047
172	TGCN	0.7	12.0	linear	IT-STGCN	2	50.0	1.110	0.037
173	TGCN	0.7	12.0	linear	STGCN	2	50.0	1.184	0.057

pd.merge(df.query("dataset=='fivenodes' and mtype=='rand'").groupby(['model','mrate','nof_filters','inter_method','method','lags','epoch'])['mse'].mean().reset_index(),
         df.query("dataset=='fivenodes' and mtype=='rand'").groupby(['model','mrate','nof_filters','inter_method','method','lags','epoch'])['mse'].std().reset_index(),
         on=['model','inter_method','method','nof_filters','mrate','lags','epoch']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate==0.8")

	model	mrate	nof_filters	inter_method	method	lags	epoch	mean	std
16	DCRNN	0.8	2.0	linear	IT-STGCN	2	50.0	1.257	0.057
17	DCRNN	0.8	2.0	linear	STGCN	2	50.0	1.255	0.040
18	DCRNN	0.8	2.0	nearest	IT-STGCN	2	50.0	1.246	0.034
19	DCRNN	0.8	2.0	nearest	STGCN	2	50.0	1.253	0.043
36	DyGrEncoder	0.8	12.0	linear	IT-STGCN	2	50.0	1.333	0.080
37	DyGrEncoder	0.8	12.0	linear	STGCN	2	50.0	1.496	0.146
38	DyGrEncoder	0.8	12.0	nearest	IT-STGCN	2	50.0	1.311	0.057
39	DyGrEncoder	0.8	12.0	nearest	STGCN	2	50.0	1.519	0.129
56	EvolveGCNH	0.8	12.0	linear	IT-STGCN	2	50.0	1.212	0.065
57	EvolveGCNH	0.8	12.0	linear	STGCN	2	50.0	1.217	0.061
58	EvolveGCNH	0.8	12.0	nearest	IT-STGCN	2	50.0	1.207	0.057
59	EvolveGCNH	0.8	12.0	nearest	STGCN	2	50.0	1.217	0.059
76	EvolveGCNO	0.8	12.0	linear	IT-STGCN	2	50.0	1.194	0.065
77	EvolveGCNO	0.8	12.0	linear	STGCN	2	50.0	1.220	0.063
78	EvolveGCNO	0.8	12.0	nearest	IT-STGCN	2	50.0	1.224	0.079
79	EvolveGCNO	0.8	12.0	nearest	STGCN	2	50.0	1.212	0.053
96	GCLSTM	0.8	4.0	linear	IT-STGCN	2	50.0	1.235	0.027
97	GCLSTM	0.8	4.0	linear	STGCN	2	50.0	1.263	0.050
98	GCLSTM	0.8	4.0	nearest	IT-STGCN	2	50.0	1.237	0.029
99	GCLSTM	0.8	4.0	nearest	STGCN	2	50.0	1.258	0.046
116	GConvGRU	0.8	12.0	linear	IT-STGCN	2	50.0	1.383	0.108
117	GConvGRU	0.8	12.0	linear	STGCN	2	50.0	2.224	0.192
118	GConvGRU	0.8	12.0	nearest	IT-STGCN	2	50.0	1.360	0.084
119	GConvGRU	0.8	12.0	nearest	STGCN	2	50.0	2.641	0.117
136	GConvLSTM	0.8	12.0	linear	IT-STGCN	2	50.0	1.298	0.060
137	GConvLSTM	0.8	12.0	linear	STGCN	2	50.0	1.442	0.111
138	GConvLSTM	0.8	12.0	nearest	IT-STGCN	2	50.0	1.312	0.065
139	GConvLSTM	0.8	12.0	nearest	STGCN	2	50.0	1.436	0.098
156	LRGCN	0.8	4.0	linear	IT-STGCN	2	50.0	1.233	0.041
157	LRGCN	0.8	4.0	linear	STGCN	2	50.0	1.250	0.038
158	LRGCN	0.8	4.0	nearest	IT-STGCN	2	50.0	1.230	0.036
159	LRGCN	0.8	4.0	nearest	STGCN	2	50.0	1.250	0.046
176	TGCN	0.8	12.0	linear	IT-STGCN	2	50.0	1.125	0.047
177	TGCN	0.8	12.0	linear	STGCN	2	50.0	1.199	0.070
178	TGCN	0.8	12.0	nearest	IT-STGCN	2	50.0	1.139	0.054
179	TGCN	0.8	12.0	nearest	STGCN	2	50.0	1.140	0.044

Block

pd.merge(df.query("dataset=='fivenodes' and mtype=='block'").groupby(['model','mrate','nof_filters','inter_method','method','epoch'])['mse'].mean().reset_index(),
         df.query("dataset=='fivenodes' and mtype=='block'").groupby(['model','mrate','nof_filters','inter_method','method','epoch'])['mse'].std().reset_index(),
         on=['model','inter_method','method','nof_filters','mrate','epoch']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	nof_filters	inter_method	method	epoch	mean	std
0	DCRNN	0.125	2.0	linear	IT-STGCN	50.0	1.232	0.033
1	DCRNN	0.125	2.0	linear	STGCN	50.0	1.260	0.051
2	DCRNN	0.125	2.0	nearest	IT-STGCN	50.0	1.222	0.025
3	DCRNN	0.125	2.0	nearest	STGCN	50.0	1.248	0.039
4	DyGrEncoder	0.125	12.0	linear	IT-STGCN	50.0	1.124	0.035
5	DyGrEncoder	0.125	12.0	linear	STGCN	50.0	1.173	0.037
6	DyGrEncoder	0.125	12.0	nearest	IT-STGCN	50.0	1.128	0.031
7	DyGrEncoder	0.125	12.0	nearest	STGCN	50.0	1.135	0.033
8	EvolveGCNH	0.125	12.0	linear	IT-STGCN	50.0	1.181	0.055
9	EvolveGCNH	0.125	12.0	linear	STGCN	50.0	1.197	0.076
10	EvolveGCNH	0.125	12.0	nearest	IT-STGCN	50.0	1.159	0.053
11	EvolveGCNH	0.125	12.0	nearest	STGCN	50.0	1.181	0.037
12	EvolveGCNO	0.125	12.0	linear	IT-STGCN	50.0	1.162	0.040
13	EvolveGCNO	0.125	12.0	linear	STGCN	50.0	1.176	0.056
14	EvolveGCNO	0.125	12.0	nearest	IT-STGCN	50.0	1.167	0.061
15	EvolveGCNO	0.125	12.0	nearest	STGCN	50.0	1.195	0.065
16	GCLSTM	0.125	4.0	linear	IT-STGCN	50.0	1.219	0.025
17	GCLSTM	0.125	4.0	linear	STGCN	50.0	1.244	0.033
18	GCLSTM	0.125	4.0	nearest	IT-STGCN	50.0	1.215	0.022
19	GCLSTM	0.125	4.0	nearest	STGCN	50.0	1.248	0.040
20	GConvGRU	0.125	12.0	linear	IT-STGCN	50.0	1.165	0.043
21	GConvGRU	0.125	12.0	linear	STGCN	50.0	1.210	0.039
22	GConvGRU	0.125	12.0	nearest	IT-STGCN	50.0	1.156	0.042
23	GConvGRU	0.125	12.0	nearest	STGCN	50.0	1.220	0.032
24	GConvLSTM	0.125	12.0	linear	IT-STGCN	50.0	1.140	0.038
25	GConvLSTM	0.125	12.0	linear	STGCN	50.0	1.172	0.055
26	GConvLSTM	0.125	12.0	nearest	IT-STGCN	50.0	1.121	0.027
27	GConvLSTM	0.125	12.0	nearest	STGCN	50.0	1.140	0.058
28	LRGCN	0.125	4.0	linear	IT-STGCN	50.0	1.220	0.020
29	LRGCN	0.125	4.0	linear	STGCN	50.0	1.251	0.037
30	LRGCN	0.125	4.0	nearest	IT-STGCN	50.0	1.216	0.030
31	LRGCN	0.125	4.0	nearest	STGCN	50.0	1.239	0.034
32	TGCN	0.125	12.0	linear	IT-STGCN	50.0	1.090	0.015
33	TGCN	0.125	12.0	linear	STGCN	50.0	1.107	0.020
34	TGCN	0.125	12.0	nearest	IT-STGCN	50.0	1.091	0.015
35	TGCN	0.125	12.0	nearest	STGCN	50.0	1.091	0.011

pd.merge(df.query("dataset=='fivenodes' and mtype=='block'").groupby(['model','mrate','nof_filters','inter_method','method','epoch'])['mse'].mean().reset_index(),
         df.query("dataset=='fivenodes' and mtype=='block'").groupby(['model','mrate','nof_filters','inter_method','method','epoch'])['mse'].std().reset_index(),
         on=['model','inter_method','method','nof_filters','mrate','epoch']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("inter_method=='linear'")

	model	mrate	nof_filters	inter_method	method	epoch	mean	std
0	DCRNN	0.125	2.0	linear	IT-STGCN	50.0	1.232	0.033
1	DCRNN	0.125	2.0	linear	STGCN	50.0	1.260	0.051
4	DyGrEncoder	0.125	12.0	linear	IT-STGCN	50.0	1.124	0.035
5	DyGrEncoder	0.125	12.0	linear	STGCN	50.0	1.173	0.037
8	EvolveGCNH	0.125	12.0	linear	IT-STGCN	50.0	1.181	0.055
9	EvolveGCNH	0.125	12.0	linear	STGCN	50.0	1.197	0.076
12	EvolveGCNO	0.125	12.0	linear	IT-STGCN	50.0	1.162	0.040
13	EvolveGCNO	0.125	12.0	linear	STGCN	50.0	1.176	0.056
16	GCLSTM	0.125	4.0	linear	IT-STGCN	50.0	1.219	0.025
17	GCLSTM	0.125	4.0	linear	STGCN	50.0	1.244	0.033
20	GConvGRU	0.125	12.0	linear	IT-STGCN	50.0	1.165	0.043
21	GConvGRU	0.125	12.0	linear	STGCN	50.0	1.210	0.039
24	GConvLSTM	0.125	12.0	linear	IT-STGCN	50.0	1.140	0.038
25	GConvLSTM	0.125	12.0	linear	STGCN	50.0	1.172	0.055
28	LRGCN	0.125	4.0	linear	IT-STGCN	50.0	1.220	0.020
29	LRGCN	0.125	4.0	linear	STGCN	50.0	1.251	0.037
32	TGCN	0.125	12.0	linear	IT-STGCN	50.0	1.090	0.015
33	TGCN	0.125	12.0	linear	STGCN	50.0	1.107	0.020

ChickenpoxDatasetLoader(lags=4)

df.query("model=='GNAR' and dataset=='chickenpox'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26518	chickenpox	GNAR	0.00000	NaN	4	NaN	NaN	NaN	1.427494	0.087380	GNAR
26519	chickenpox	GNAR	0.30000	rand	4	NaN	linear	NaN	1.427494	0.086311	GNAR
26520	chickenpox	GNAR	0.80000	rand	4	NaN	linear	NaN	1.427494	0.101606	GNAR
26521	chickenpox	GNAR	0.00000	NaN	4	NaN	NaN	NaN	1.427494	0.100545	GNAR
26522	chickenpox	GNAR	0.30000	rand	4	NaN	linear	NaN	1.427494	0.084384	GNAR
26523	chickenpox	GNAR	0.80000	rand	4	NaN	linear	NaN	1.427494	0.095799	GNAR
26524	chickenpox	GNAR	0.00000	NaN	4	NaN	NaN	NaN	1.427494	0.135489	GNAR
26525	chickenpox	GNAR	0.30000	rand	4	NaN	linear	NaN	1.427494	0.097362	GNAR
26526	chickenpox	GNAR	0.80000	rand	4	NaN	linear	NaN	1.427494	0.108125	GNAR
26566	chickenpox	GNAR	0.28777	block	4	NaN	linear	NaN	1.427494	0.069579	GNAR
26567	chickenpox	GNAR	0.28777	block	4	NaN	nearest	NaN	1.427494	0.073124	GNAR
26568	chickenpox	GNAR	0.28777	block	4	NaN	linear	NaN	1.427494	0.070857	GNAR
26569	chickenpox	GNAR	0.28777	block	4	NaN	nearest	NaN	1.427494	0.066239	GNAR
26570	chickenpox	GNAR	0.28777	block	4	NaN	linear	NaN	1.427494	0.070568	GNAR
26571	chickenpox	GNAR	0.28777	block	4	NaN	nearest	NaN	1.427494	0.070212	GNAR
26773	chickenpox	GNAR	0.50000	rand	4	NaN	linear	NaN	1.427494	0.077680	GNAR
26774	chickenpox	GNAR	0.60000	rand	4	NaN	linear	NaN	1.427494	0.104648	GNAR
26775	chickenpox	GNAR	0.50000	rand	4	NaN	linear	NaN	1.427494	0.123352	GNAR
26776	chickenpox	GNAR	0.60000	rand	4	NaN	linear	NaN	1.427494	0.073757	GNAR
26777	chickenpox	GNAR	0.50000	rand	4	NaN	linear	NaN	1.427494	0.082407	GNAR
26778	chickenpox	GNAR	0.60000	rand	4	NaN	linear	NaN	1.427494	0.073220	GNAR

Baseline

pd.merge(df.query("dataset=='chickenpox' and mtype!='rand' and mtype!='block'").groupby(['model','nof_filters'])['mse'].mean().reset_index(),
         df.query("dataset=='chickenpox' and mtype!='rand' and mtype!='block'").groupby(['model','nof_filters'])['mse'].std().reset_index(),
         on=['model','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	nof_filters	mean	std
0	DCRNN	16.0	0.727	0.009
1	DyGrEncoder	12.0	0.906	0.051
2	EvolveGCNH	32.0	1.000	0.020
3	EvolveGCNO	32.0	0.986	0.018
4	GCLSTM	16.0	0.885	0.051
5	GConvGRU	16.0	0.752	0.013
6	GConvLSTM	32.0	0.959	0.088
7	LRGCN	8.0	0.868	0.047
8	TGCN	12.0	1.090	0.042

Random

fig, ax = plt.subplots(3, 3,figsize=(40,20))

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==16 and lags==4 and epoch==50 and model=='GConvGRU' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,0],grid=False,widths=0.5)
ax[0,0].set_title('GConvGRU')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==32 and lags==4 and epoch==50 and model=='GConvLSTM' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,1],grid=False,widths=0.5)
ax[0,1].set_title('GConvLSTM')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==16 and lags==4 and epoch==50 and model=='GCLSTM' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,2],grid=False,widths=0.5)
ax[0,2].set_title('GCLSTM')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==8 and lags==4 and epoch==50 and model=='LRGCN' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,0],grid=False,widths=0.5)
ax[1,0].set_title('LRGCN')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==4 and epoch==50 and model=='DyGrEncoder' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,1],grid=False,widths=0.5)
ax[1,1].set_title('DyGrEncoder')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and lags==4 and epoch==50 and model=='EvolveGCNH' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,2],grid=False,widths=0.5)
ax[1,2].set_title('EvolveGCNH')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and lags==4 and epoch==50 and model=='EvolveGCNO' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,0],grid=False,widths=0.5)
ax[2,0].set_title('EvolveGCNO')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==4 and epoch==50 and model=='TGCN' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,1],grid=False,widths=0.5)
ax[2,1].set_title('TGCN')

df.query("dataset=='chickenpox' and mtype=='rand' and inter_method == 'linear' and nof_filters==16 and lags==4 and epoch==50 and model=='DCRNN' and mrate in [ 0.3   , 0.5   ,   0.6  , 0.8 ]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,2],grid=False,widths=0.5)
ax[2,2].set_title('DCRNN')


for ax in ax.flat:
    ax.set_yticklabels([])
    ax.set_yscale('log')
    ax.axvline(x=2.5, color='black', linestyle='-')
    ax.axvline(x=4.5, color='black', linestyle='-')
    ax.axvline(x=6.5, color='black', linestyle='-')
    ax.axvline(x=8.5, color='black', linestyle='-')
    ax.set_xticklabels(['IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN'])
    ax.set_xlabel('')
    ax.set_ylabel('')
    
fig.suptitle('',fontsize=40)

Text(0.5, 0.98, '')

pd.merge(df.query("dataset=='chickenpox' and mtype=='rand'").groupby(['model','mrate','inter_method','nof_filters','method'])['mse'].mean().reset_index(),
         df.query("dataset=='chickenpox' and mtype=='rand'").groupby(['model','mrate','inter_method','nof_filters','method'])['mse'].std().reset_index(),
         on=['model','method','inter_method','mrate','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate==0.3")

	model	mrate	inter_method	nof_filters	method	mean	std
0	DCRNN	0.3	linear	16.0	IT-STGCN	0.797	0.010
1	DCRNN	0.3	linear	16.0	STGCN	1.032	0.039
8	DyGrEncoder	0.3	linear	12.0	IT-STGCN	0.868	0.028
9	DyGrEncoder	0.3	linear	12.0	STGCN	1.080	0.037
16	EvolveGCNH	0.3	linear	32.0	IT-STGCN	1.011	0.019
17	EvolveGCNH	0.3	linear	32.0	STGCN	1.058	0.015
24	EvolveGCNO	0.3	linear	32.0	IT-STGCN	0.998	0.019
25	EvolveGCNO	0.3	linear	32.0	STGCN	1.054	0.011
32	GCLSTM	0.3	linear	16.0	IT-STGCN	0.850	0.022
33	GCLSTM	0.3	linear	16.0	STGCN	1.050	0.036
40	GConvGRU	0.3	linear	16.0	IT-STGCN	0.851	0.031
41	GConvGRU	0.3	linear	16.0	STGCN	1.087	0.046
48	GConvLSTM	0.3	linear	32.0	IT-STGCN	0.872	0.035
49	GConvLSTM	0.3	linear	32.0	STGCN	1.114	0.057
56	LRGCN	0.3	linear	8.0	IT-STGCN	0.870	0.035
57	LRGCN	0.3	linear	8.0	STGCN	1.086	0.029
64	TGCN	0.3	linear	12.0	IT-STGCN	1.042	0.020
65	TGCN	0.3	linear	12.0	STGCN	1.054	0.015

pd.merge(df.query("dataset=='chickenpox' and mtype=='rand'").groupby(['model','mrate','inter_method','nof_filters','method'])['mse'].mean().reset_index(),
         df.query("dataset=='chickenpox' and mtype=='rand'").groupby(['model','mrate','inter_method','nof_filters','method'])['mse'].std().reset_index(),
         on=['model','method','inter_method','mrate','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate!=0.3")

	model	mrate	inter_method	nof_filters	method	mean	std
2	DCRNN	0.5	linear	16.0	IT-STGCN	0.869	0.018
3	DCRNN	0.5	linear	16.0	STGCN	1.473	0.058
4	DCRNN	0.6	linear	16.0	IT-STGCN	0.973	0.032
5	DCRNN	0.6	linear	16.0	STGCN	1.848	0.072
6	DCRNN	0.8	linear	16.0	IT-STGCN	1.467	0.076
7	DCRNN	0.8	linear	16.0	STGCN	2.287	0.074
10	DyGrEncoder	0.5	linear	12.0	IT-STGCN	0.915	0.029
11	DyGrEncoder	0.5	linear	12.0	STGCN	1.540	0.045
12	DyGrEncoder	0.6	linear	12.0	IT-STGCN	1.013	0.035
13	DyGrEncoder	0.6	linear	12.0	STGCN	1.807	0.068
14	DyGrEncoder	0.8	linear	12.0	IT-STGCN	1.399	0.063
15	DyGrEncoder	0.8	linear	12.0	STGCN	2.127	0.240
18	EvolveGCNH	0.5	linear	16.0	IT-STGCN	1.026	0.017
19	EvolveGCNH	0.5	linear	16.0	STGCN	1.122	0.035
20	EvolveGCNH	0.6	linear	16.0	IT-STGCN	1.054	0.024
21	EvolveGCNH	0.6	linear	16.0	STGCN	1.162	0.044
22	EvolveGCNH	0.8	linear	32.0	IT-STGCN	1.140	0.042
23	EvolveGCNH	0.8	linear	32.0	STGCN	1.203	0.061
26	EvolveGCNO	0.5	linear	16.0	IT-STGCN	1.020	0.014
27	EvolveGCNO	0.5	linear	16.0	STGCN	1.145	0.027
28	EvolveGCNO	0.6	linear	16.0	IT-STGCN	1.050	0.018
29	EvolveGCNO	0.6	linear	16.0	STGCN	1.205	0.055
30	EvolveGCNO	0.8	linear	32.0	IT-STGCN	1.161	0.054
31	EvolveGCNO	0.8	linear	32.0	STGCN	1.234	0.096
34	GCLSTM	0.5	linear	16.0	IT-STGCN	0.899	0.023
35	GCLSTM	0.5	linear	16.0	STGCN	1.514	0.050
36	GCLSTM	0.6	linear	16.0	IT-STGCN	1.003	0.030
37	GCLSTM	0.6	linear	16.0	STGCN	1.808	0.069
38	GCLSTM	0.8	linear	16.0	IT-STGCN	1.371	0.072
39	GCLSTM	0.8	linear	16.0	STGCN	2.172	0.186
42	GConvGRU	0.5	linear	16.0	IT-STGCN	0.958	0.072
43	GConvGRU	0.5	linear	16.0	STGCN	1.530	0.106
44	GConvGRU	0.6	linear	16.0	IT-STGCN	1.120	0.072
45	GConvGRU	0.6	linear	16.0	STGCN	1.753	0.181
46	GConvGRU	0.8	linear	16.0	IT-STGCN	1.586	0.199
47	GConvGRU	0.8	linear	16.0	STGCN	2.529	0.292
50	GConvLSTM	0.5	linear	32.0	IT-STGCN	0.901	0.024
51	GConvLSTM	0.5	linear	32.0	STGCN	1.518	0.063
52	GConvLSTM	0.6	linear	32.0	IT-STGCN	1.004	0.038
53	GConvLSTM	0.6	linear	32.0	STGCN	1.787	0.055
54	GConvLSTM	0.8	linear	32.0	IT-STGCN	1.433	0.080
55	GConvLSTM	0.8	linear	32.0	STGCN	2.522	0.111
58	LRGCN	0.5	linear	8.0	IT-STGCN	0.931	0.034
59	LRGCN	0.5	linear	8.0	STGCN	1.458	0.068
60	LRGCN	0.6	linear	8.0	IT-STGCN	1.017	0.029
61	LRGCN	0.6	linear	8.0	STGCN	1.615	0.134
62	LRGCN	0.8	linear	8.0	IT-STGCN	1.334	0.071
63	LRGCN	0.8	linear	8.0	STGCN	1.632	0.156
66	TGCN	0.5	linear	12.0	IT-STGCN	1.032	0.012
67	TGCN	0.5	linear	12.0	STGCN	1.167	0.018
68	TGCN	0.6	linear	12.0	IT-STGCN	1.054	0.014
69	TGCN	0.6	linear	12.0	STGCN	1.242	0.021
70	TGCN	0.8	linear	12.0	IT-STGCN	1.183	0.028
71	TGCN	0.8	linear	12.0	STGCN	1.466	0.064

Block

pd.merge(df.query("dataset=='chickenpox' and mtype=='block'").groupby(['model','inter_method','mrate','nof_filters','method'])['mse'].mean().reset_index(),
         df.query("dataset=='chickenpox' and mtype=='block'").groupby(['model','inter_method','mrate','nof_filters','method'])['mse'].std().reset_index(),
         on=['model','method','inter_method','mrate','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	inter_method	mrate	nof_filters	method	mean	std
0	DCRNN	linear	0.288	16.0	IT-STGCN	0.740	0.007
1	DCRNN	linear	0.288	16.0	STGCN	0.812	0.006
2	DCRNN	nearest	0.288	16.0	IT-STGCN	0.738	0.007
3	DCRNN	nearest	0.288	16.0	STGCN	0.832	0.009
4	DyGrEncoder	linear	0.288	12.0	IT-STGCN	0.899	0.035
5	DyGrEncoder	linear	0.288	12.0	STGCN	0.912	0.043
6	DyGrEncoder	nearest	0.288	12.0	IT-STGCN	0.909	0.043
7	DyGrEncoder	nearest	0.288	12.0	STGCN	0.930	0.035
8	EvolveGCNH	linear	0.288	32.0	IT-STGCN	1.007	0.021
9	EvolveGCNH	linear	0.288	32.0	STGCN	1.027	0.023
10	EvolveGCNH	nearest	0.288	32.0	IT-STGCN	1.011	0.018
11	EvolveGCNH	nearest	0.288	32.0	STGCN	1.030	0.017
12	EvolveGCNO	linear	0.288	32.0	IT-STGCN	1.002	0.015
13	EvolveGCNO	linear	0.288	32.0	STGCN	1.028	0.016
14	EvolveGCNO	nearest	0.288	32.0	IT-STGCN	0.999	0.022
15	EvolveGCNO	nearest	0.288	32.0	STGCN	1.026	0.015
16	GCLSTM	linear	0.288	16.0	IT-STGCN	0.883	0.045
17	GCLSTM	linear	0.288	16.0	STGCN	0.890	0.033
18	GCLSTM	nearest	0.288	16.0	IT-STGCN	0.901	0.054
19	GCLSTM	nearest	0.288	16.0	STGCN	0.885	0.042
20	GConvGRU	linear	0.288	16.0	IT-STGCN	0.807	0.016
21	GConvGRU	linear	0.288	16.0	STGCN	0.828	0.022
22	GConvGRU	nearest	0.288	16.0	IT-STGCN	0.824	0.023
23	GConvGRU	nearest	0.288	16.0	STGCN	0.828	0.022
24	GConvLSTM	linear	0.288	32.0	IT-STGCN	0.911	0.069
25	GConvLSTM	linear	0.288	32.0	STGCN	0.900	0.049
26	GConvLSTM	nearest	0.288	32.0	IT-STGCN	0.885	0.040
27	GConvLSTM	nearest	0.288	32.0	STGCN	0.896	0.054
28	LRGCN	linear	0.288	8.0	IT-STGCN	0.888	0.035
29	LRGCN	linear	0.288	8.0	STGCN	0.911	0.047
30	LRGCN	nearest	0.288	8.0	IT-STGCN	0.888	0.041
31	LRGCN	nearest	0.288	8.0	STGCN	0.902	0.039
32	TGCN	linear	0.288	12.0	IT-STGCN	1.065	0.031
33	TGCN	linear	0.288	12.0	STGCN	1.082	0.028
34	TGCN	nearest	0.288	12.0	IT-STGCN	1.070	0.028
35	TGCN	nearest	0.288	12.0	STGCN	1.079	0.027

pd.merge(df.query("dataset=='chickenpox' and mtype=='block'").groupby(['model','inter_method','mrate','nof_filters','method'])['mse'].mean().reset_index(),
         df.query("dataset=='chickenpox' and mtype=='block'").groupby(['model','inter_method','mrate','nof_filters','method'])['mse'].std().reset_index(),
         on=['model','method','inter_method','mrate','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("inter_method=='linear'")

	model	inter_method	mrate	nof_filters	method	mean	std
0	DCRNN	linear	0.288	16.0	IT-STGCN	0.740	0.007
1	DCRNN	linear	0.288	16.0	STGCN	0.812	0.006
4	DyGrEncoder	linear	0.288	12.0	IT-STGCN	0.899	0.035
5	DyGrEncoder	linear	0.288	12.0	STGCN	0.912	0.043
8	EvolveGCNH	linear	0.288	32.0	IT-STGCN	1.007	0.021
9	EvolveGCNH	linear	0.288	32.0	STGCN	1.027	0.023
12	EvolveGCNO	linear	0.288	32.0	IT-STGCN	1.002	0.015
13	EvolveGCNO	linear	0.288	32.0	STGCN	1.028	0.016
16	GCLSTM	linear	0.288	16.0	IT-STGCN	0.883	0.045
17	GCLSTM	linear	0.288	16.0	STGCN	0.890	0.033
20	GConvGRU	linear	0.288	16.0	IT-STGCN	0.807	0.016
21	GConvGRU	linear	0.288	16.0	STGCN	0.828	0.022
24	GConvLSTM	linear	0.288	32.0	IT-STGCN	0.911	0.069
25	GConvLSTM	linear	0.288	32.0	STGCN	0.900	0.049
28	LRGCN	linear	0.288	8.0	IT-STGCN	0.888	0.035
29	LRGCN	linear	0.288	8.0	STGCN	0.911	0.047
32	TGCN	linear	0.288	12.0	IT-STGCN	1.065	0.031
33	TGCN	linear	0.288	12.0	STGCN	1.082	0.028

PedalMeDatasetLoader (lags=4)

df.query("model=='GNAR' and dataset=='pedalme'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26527	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.031931	GNAR
26528	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.027079	GNAR
26529	pedalme	GNAR	0.300000	rand	4	NaN	linear	NaN	1.302679	0.036771	GNAR
26530	pedalme	GNAR	0.300000	rand	4	NaN	nearest	NaN	1.302679	0.022349	GNAR
26531	pedalme	GNAR	0.600000	rand	4	NaN	linear	NaN	1.302679	0.022278	GNAR
26532	pedalme	GNAR	0.600000	rand	4	NaN	nearest	NaN	1.302679	0.029977	GNAR
26533	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.022007	GNAR
26534	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.026209	GNAR
26535	pedalme	GNAR	0.300000	rand	4	NaN	linear	NaN	1.302679	0.024794	GNAR
26536	pedalme	GNAR	0.300000	rand	4	NaN	nearest	NaN	1.302679	0.020568	GNAR
26537	pedalme	GNAR	0.600000	rand	4	NaN	linear	NaN	1.302679	0.022407	GNAR
26538	pedalme	GNAR	0.600000	rand	4	NaN	nearest	NaN	1.302679	0.021174	GNAR
26539	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.035516	GNAR
26540	pedalme	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.302679	0.020410	GNAR
26541	pedalme	GNAR	0.300000	rand	4	NaN	linear	NaN	1.302679	0.035666	GNAR
26542	pedalme	GNAR	0.300000	rand	4	NaN	nearest	NaN	1.302679	0.021164	GNAR
26543	pedalme	GNAR	0.600000	rand	4	NaN	linear	NaN	1.302679	0.020147	GNAR
26544	pedalme	GNAR	0.600000	rand	4	NaN	nearest	NaN	1.302679	0.021981	GNAR
26572	pedalme	GNAR	0.285714	block	4	NaN	linear	NaN	1.302679	0.017945	GNAR
26573	pedalme	GNAR	0.285714	block	4	NaN	nearest	NaN	1.302679	0.021973	GNAR
26574	pedalme	GNAR	0.285714	block	4	NaN	linear	NaN	1.302679	0.023076	GNAR
26575	pedalme	GNAR	0.285714	block	4	NaN	nearest	NaN	1.302679	0.022257	GNAR
26576	pedalme	GNAR	0.285714	block	4	NaN	linear	NaN	1.302679	0.023361	GNAR
26577	pedalme	GNAR	0.285714	block	4	NaN	nearest	NaN	1.302679	0.022614	GNAR
26779	pedalme	GNAR	0.500000	rand	4	NaN	linear	NaN	1.302679	0.038041	GNAR
26780	pedalme	GNAR	0.500000	rand	4	NaN	nearest	NaN	1.302679	0.022290	GNAR
26781	pedalme	GNAR	0.800000	rand	4	NaN	linear	NaN	1.302679	0.019467	GNAR
26782	pedalme	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.302679	0.018487	GNAR
26783	pedalme	GNAR	0.500000	rand	4	NaN	linear	NaN	1.302679	0.022977	GNAR
26784	pedalme	GNAR	0.500000	rand	4	NaN	nearest	NaN	1.302679	0.039169	GNAR
26785	pedalme	GNAR	0.800000	rand	4	NaN	linear	NaN	1.302679	0.023175	GNAR
26786	pedalme	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.302679	0.023105	GNAR
26787	pedalme	GNAR	0.500000	rand	4	NaN	linear	NaN	1.302679	0.022839	GNAR
26788	pedalme	GNAR	0.500000	rand	4	NaN	nearest	NaN	1.302679	0.023041	GNAR
26789	pedalme	GNAR	0.800000	rand	4	NaN	linear	NaN	1.302679	0.060872	GNAR
26790	pedalme	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.302679	0.023526	GNAR

Baseline

pd.merge(df.query("dataset=='pedalme' and mtype!='rand' and mtype!='block'").groupby(['model','lags','nof_filters'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype!='rand' and mtype!='block'").groupby(['model','lags','nof_filters'])['mse'].std().reset_index(),
         on=['model','lags','nof_filters']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4")

	model	lags	nof_filters	mean	std
0	DCRNN	4	8.0	1.131	0.015
1	DyGrEncoder	4	12.0	1.190	0.047
2	EvolveGCNH	4	2.0	1.213	0.057
3	EvolveGCNO	4	2.0	1.223	0.051
4	GCLSTM	4	4.0	1.181	0.040
5	GConvGRU	4	12.0	1.233	0.107
6	GConvLSTM	4	2.0	1.214	0.055
7	LRGCN	4	8.0	1.191	0.054
8	TGCN	4	12.0	1.307	0.075

Random

fig, ax = plt.subplots(3, 3,figsize=(40,20))

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==4 and epoch==50 and model=='GConvGRU' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,0],grid=False,widths=0.5)
ax[0,0].set_title('GConvGRU')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==2 and lags==4 and epoch==50 and model=='GConvLSTM' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,1],grid=False,widths=0.5)
ax[0,1].set_title('GConvLSTM')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==4 and lags==4 and epoch==50 and model=='GCLSTM' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,2],grid=False,widths=0.5)
ax[0,2].set_title('GCLSTM')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==8 and lags==4 and epoch==50 and model=='LRGCN' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,0],grid=False,widths=0.5)
ax[1,0].set_title('LRGCN')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==4 and epoch==50 and model=='DyGrEncoder' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,1],grid=False,widths=0.5)
ax[1,1].set_title('DyGrEncoder')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and lags==4 and epoch==50 and model=='EvolveGCNH' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,2],grid=False,widths=0.5)
ax[1,2].set_title('EvolveGCNH')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and lags==4 and epoch==50 and model=='EvolveGCNO' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,0],grid=False,widths=0.5)
ax[2,0].set_title('EvolveGCNO')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==4 and epoch==50 and model=='TGCN' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,1],grid=False,widths=0.5)
ax[2,1].set_title('TGCN')

df.query("dataset=='pedalme' and mtype=='rand' and inter_method == 'linear' and nof_filters==8 and lags==4 and epoch==50 and model=='DCRNN' and mrate in [0.3,0.5,0.6,0.8]").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,2],grid=False,widths=0.5)
ax[2,2].set_title('DCRNN')


for ax in ax.flat:
    ax.set_yticklabels([])
    ax.set_yscale('log')
    ax.axvline(x=2.5, color='black', linestyle='-')
    ax.axvline(x=4.5, color='black', linestyle='-')
    ax.axvline(x=6.5, color='black', linestyle='-')
    ax.axvline(x=8.5, color='black', linestyle='-')
    ax.set_xticklabels(['IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN'])
    ax.set_xlabel('')
    ax.set_ylabel('')
    
fig.suptitle('',fontsize=40)

Text(0.5, 0.98, '')

pd.merge(df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','nof_filters','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate == 0.3")

	model	mrate	lags	nof_filters	inter_method	method	mean	std
0	DCRNN	0.3	4	8.0	linear	IT-STGCN	1.190	0.029
1	DCRNN	0.3	4	8.0	linear	STGCN	1.277	0.064
2	DCRNN	0.3	4	8.0	nearest	IT-STGCN	1.179	0.035
3	DCRNN	0.3	4	8.0	nearest	STGCN	1.278	0.060
16	DyGrEncoder	0.3	4	12.0	linear	IT-STGCN	1.207	0.046
17	DyGrEncoder	0.3	4	12.0	linear	STGCN	1.279	0.061
18	DyGrEncoder	0.3	4	12.0	nearest	IT-STGCN	1.205	0.075
19	DyGrEncoder	0.3	4	12.0	nearest	STGCN	1.289	0.096
32	EvolveGCNH	0.3	4	2.0	linear	IT-STGCN	1.245	0.069
33	EvolveGCNH	0.3	4	2.0	linear	STGCN	1.273	0.057
34	EvolveGCNH	0.3	4	2.0	nearest	IT-STGCN	1.238	0.046
35	EvolveGCNH	0.3	4	2.0	nearest	STGCN	1.244	0.054
48	EvolveGCNO	0.3	4	2.0	linear	IT-STGCN	1.251	0.072
49	EvolveGCNO	0.3	4	2.0	linear	STGCN	1.267	0.072
50	EvolveGCNO	0.3	4	2.0	nearest	IT-STGCN	1.251	0.057
51	EvolveGCNO	0.3	4	2.0	nearest	STGCN	1.265	0.056
64	GCLSTM	0.3	4	4.0	linear	IT-STGCN	1.202	0.029
65	GCLSTM	0.3	4	4.0	linear	STGCN	1.267	0.041
66	GCLSTM	0.3	4	4.0	nearest	IT-STGCN	1.211	0.039
67	GCLSTM	0.3	4	4.0	nearest	STGCN	1.256	0.038
80	GConvGRU	0.3	4	12.0	linear	IT-STGCN	1.354	0.134
81	GConvGRU	0.3	4	12.0	linear	STGCN	1.575	0.198
82	GConvGRU	0.3	4	12.0	nearest	IT-STGCN	1.385	0.173
83	GConvGRU	0.3	4	12.0	nearest	STGCN	1.527	0.342
96	GConvLSTM	0.3	4	2.0	linear	IT-STGCN	1.227	0.056
97	GConvLSTM	0.3	4	2.0	linear	STGCN	1.244	0.041
98	GConvLSTM	0.3	4	2.0	nearest	IT-STGCN	1.224	0.035
99	GConvLSTM	0.3	4	2.0	nearest	STGCN	1.266	0.068
112	LRGCN	0.3	4	8.0	linear	IT-STGCN	1.210	0.039
113	LRGCN	0.3	4	8.0	linear	STGCN	1.274	0.045
114	LRGCN	0.3	4	8.0	nearest	IT-STGCN	1.220	0.046
115	LRGCN	0.3	4	8.0	nearest	STGCN	1.287	0.065
128	TGCN	0.3	4	12.0	linear	IT-STGCN	1.280	0.070
129	TGCN	0.3	4	12.0	linear	STGCN	1.302	0.112
130	TGCN	0.3	4	12.0	nearest	IT-STGCN	1.248	0.074
131	TGCN	0.3	4	12.0	nearest	STGCN	1.291	0.111

pd.merge(df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','nof_filters','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate != 0.3")

	model	mrate	lags	nof_filters	inter_method	method	mean	std
4	DCRNN	0.5	4	8.0	linear	IT-STGCN	1.238	0.080
5	DCRNN	0.5	4	8.0	linear	STGCN	1.451	0.061
6	DCRNN	0.5	4	8.0	nearest	IT-STGCN	1.232	0.061
7	DCRNN	0.5	4	8.0	nearest	STGCN	1.447	0.073
8	DCRNN	0.6	4	8.0	linear	IT-STGCN	1.314	0.072
...	...	...	...	...	...	...	...	...
139	TGCN	0.6	4	12.0	nearest	STGCN	1.301	0.090
140	TGCN	0.8	4	12.0	linear	IT-STGCN	1.294	0.063
141	TGCN	0.8	4	12.0	linear	STGCN	1.289	0.065
142	TGCN	0.8	4	12.0	nearest	IT-STGCN	1.258	0.053
143	TGCN	0.8	4	12.0	nearest	STGCN	1.277	0.080

108 rows × 8 columns

pd.merge(df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate != 0.3").\
query("inter_method=='nearest'")

	model	mrate	lags	inter_method	method	mean	std
6	DCRNN	0.5	4	nearest	IT-STGCN	1.232	0.061
7	DCRNN	0.5	4	nearest	STGCN	1.447	0.073
10	DCRNN	0.6	4	nearest	IT-STGCN	1.303	0.078
11	DCRNN	0.6	4	nearest	STGCN	1.509	0.068
14	DCRNN	0.8	4	nearest	IT-STGCN	1.527	0.079
15	DCRNN	0.8	4	nearest	STGCN	1.616	0.075
22	DyGrEncoder	0.5	4	nearest	IT-STGCN	1.236	0.059
23	DyGrEncoder	0.5	4	nearest	STGCN	1.427	0.076
26	DyGrEncoder	0.6	4	nearest	IT-STGCN	1.285	0.051
27	DyGrEncoder	0.6	4	nearest	STGCN	1.513	0.083
30	DyGrEncoder	0.8	4	nearest	IT-STGCN	1.476	0.098
31	DyGrEncoder	0.8	4	nearest	STGCN	1.614	0.096
38	EvolveGCNH	0.5	4	nearest	IT-STGCN	1.242	0.058
39	EvolveGCNH	0.5	4	nearest	STGCN	1.285	0.049
42	EvolveGCNH	0.6	4	nearest	IT-STGCN	1.262	0.091
43	EvolveGCNH	0.6	4	nearest	STGCN	1.284	0.066
46	EvolveGCNH	0.8	4	nearest	IT-STGCN	1.299	0.082
47	EvolveGCNH	0.8	4	nearest	STGCN	1.292	0.074
54	EvolveGCNO	0.5	4	nearest	IT-STGCN	1.254	0.037
55	EvolveGCNO	0.5	4	nearest	STGCN	1.274	0.080
58	EvolveGCNO	0.6	4	nearest	IT-STGCN	1.267	0.067
59	EvolveGCNO	0.6	4	nearest	STGCN	1.292	0.075
62	EvolveGCNO	0.8	4	nearest	IT-STGCN	1.312	0.073
63	EvolveGCNO	0.8	4	nearest	STGCN	1.334	0.144
70	GCLSTM	0.5	4	nearest	IT-STGCN	1.221	0.042
71	GCLSTM	0.5	4	nearest	STGCN	1.333	0.066
74	GCLSTM	0.6	4	nearest	IT-STGCN	1.259	0.042
75	GCLSTM	0.6	4	nearest	STGCN	1.365	0.064
78	GCLSTM	0.8	4	nearest	IT-STGCN	1.352	0.050
79	GCLSTM	0.8	4	nearest	STGCN	1.396	0.079
86	GConvGRU	0.5	4	nearest	IT-STGCN	1.507	0.235
87	GConvGRU	0.5	4	nearest	STGCN	1.673	0.223
90	GConvGRU	0.6	4	nearest	IT-STGCN	1.625	0.324
91	GConvGRU	0.6	4	nearest	STGCN	1.851	0.254
94	GConvGRU	0.8	4	nearest	IT-STGCN	1.608	0.243
95	GConvGRU	0.8	4	nearest	STGCN	1.871	0.214
102	GConvLSTM	0.5	4	nearest	IT-STGCN	1.235	0.042
103	GConvLSTM	0.5	4	nearest	STGCN	1.283	0.071
106	GConvLSTM	0.6	4	nearest	IT-STGCN	1.248	0.045
107	GConvLSTM	0.6	4	nearest	STGCN	1.274	0.078
110	GConvLSTM	0.8	4	nearest	IT-STGCN	1.287	0.065
111	GConvLSTM	0.8	4	nearest	STGCN	1.364	0.069
115	GNAR	0.5	4	nearest	GNAR	1.303	0.000
117	GNAR	0.6	4	nearest	GNAR	1.303	0.000
119	GNAR	0.8	4	nearest	GNAR	1.303	0.000
126	LRGCN	0.5	4	nearest	IT-STGCN	1.240	0.036
127	LRGCN	0.5	4	nearest	STGCN	1.379	0.079
130	LRGCN	0.6	4	nearest	IT-STGCN	1.286	0.033
131	LRGCN	0.6	4	nearest	STGCN	1.462	0.084
134	LRGCN	0.8	4	nearest	IT-STGCN	1.436	0.091
135	LRGCN	0.8	4	nearest	STGCN	1.529	0.071
142	TGCN	0.5	4	nearest	IT-STGCN	1.259	0.066
143	TGCN	0.5	4	nearest	STGCN	1.287	0.065
146	TGCN	0.6	4	nearest	IT-STGCN	1.260	0.072
147	TGCN	0.6	4	nearest	STGCN	1.301	0.090
150	TGCN	0.8	4	nearest	IT-STGCN	1.258	0.053
151	TGCN	0.8	4	nearest	STGCN	1.277	0.080

Block

pd.merge(df.query("dataset=='pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4")

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.286	4	linear	IT-STGCN	1.154	0.014
1	DCRNN	0.286	4	linear	STGCN	1.248	0.019
2	DCRNN	0.286	4	nearest	IT-STGCN	1.150	0.014
3	DCRNN	0.286	4	nearest	STGCN	1.304	0.021
4	DyGrEncoder	0.286	4	linear	IT-STGCN	1.167	0.040
5	DyGrEncoder	0.286	4	linear	STGCN	1.222	0.054
6	DyGrEncoder	0.286	4	nearest	IT-STGCN	1.165	0.032
7	DyGrEncoder	0.286	4	nearest	STGCN	1.269	0.066
8	EvolveGCNH	0.286	4	linear	IT-STGCN	1.259	0.085
9	EvolveGCNH	0.286	4	linear	STGCN	1.246	0.073
10	EvolveGCNH	0.286	4	nearest	IT-STGCN	1.222	0.040
11	EvolveGCNH	0.286	4	nearest	STGCN	1.265	0.072
12	EvolveGCNO	0.286	4	linear	IT-STGCN	1.246	0.034
13	EvolveGCNO	0.286	4	linear	STGCN	1.230	0.056
14	EvolveGCNO	0.286	4	nearest	IT-STGCN	1.245	0.045
15	EvolveGCNO	0.286	4	nearest	STGCN	1.246	0.035
16	GCLSTM	0.286	4	linear	IT-STGCN	1.182	0.031
17	GCLSTM	0.286	4	linear	STGCN	1.211	0.023
18	GCLSTM	0.286	4	nearest	IT-STGCN	1.195	0.029
19	GCLSTM	0.286	4	nearest	STGCN	1.248	0.019
20	GConvGRU	0.286	4	linear	IT-STGCN	1.329	0.131
21	GConvGRU	0.286	4	linear	STGCN	1.320	0.111
22	GConvGRU	0.286	4	nearest	IT-STGCN	1.289	0.115
23	GConvGRU	0.286	4	nearest	STGCN	1.270	0.114
24	GConvLSTM	0.286	4	linear	IT-STGCN	1.241	0.069
25	GConvLSTM	0.286	4	linear	STGCN	1.223	0.042
26	GConvLSTM	0.286	4	nearest	IT-STGCN	1.222	0.039
27	GConvLSTM	0.286	4	nearest	STGCN	1.237	0.046
28	GNAR	0.286	4	linear	GNAR	1.303	0.000
29	GNAR	0.286	4	nearest	GNAR	1.303	0.000
30	LRGCN	0.286	4	linear	IT-STGCN	1.169	0.040
31	LRGCN	0.286	4	linear	STGCN	1.204	0.032
32	LRGCN	0.286	4	nearest	IT-STGCN	1.165	0.035
33	LRGCN	0.286	4	nearest	STGCN	1.263	0.033
34	TGCN	0.286	4	linear	IT-STGCN	1.278	0.056
35	TGCN	0.286	4	linear	STGCN	1.244	0.071
36	TGCN	0.286	4	nearest	IT-STGCN	1.262	0.066
37	TGCN	0.286	4	nearest	STGCN	1.232	0.069

pd.merge(df.query("dataset=='pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).\
query("lags==4 and inter_method=='nearest'")

	model	mrate	lags	inter_method	method	mean	std
2	DCRNN	0.286	4	nearest	IT-STGCN	1.150	0.014
3	DCRNN	0.286	4	nearest	STGCN	1.304	0.021
6	DyGrEncoder	0.286	4	nearest	IT-STGCN	1.165	0.032
7	DyGrEncoder	0.286	4	nearest	STGCN	1.269	0.066
10	EvolveGCNH	0.286	4	nearest	IT-STGCN	1.222	0.040
11	EvolveGCNH	0.286	4	nearest	STGCN	1.265	0.072
14	EvolveGCNO	0.286	4	nearest	IT-STGCN	1.245	0.045
15	EvolveGCNO	0.286	4	nearest	STGCN	1.246	0.035
18	GCLSTM	0.286	4	nearest	IT-STGCN	1.195	0.029
19	GCLSTM	0.286	4	nearest	STGCN	1.248	0.019
22	GConvGRU	0.286	4	nearest	IT-STGCN	1.289	0.115
23	GConvGRU	0.286	4	nearest	STGCN	1.270	0.114
26	GConvLSTM	0.286	4	nearest	IT-STGCN	1.222	0.039
27	GConvLSTM	0.286	4	nearest	STGCN	1.237	0.046
29	GNAR	0.286	4	nearest	GNAR	1.303	0.000
32	LRGCN	0.286	4	nearest	IT-STGCN	1.165	0.035
33	LRGCN	0.286	4	nearest	STGCN	1.263	0.033
36	TGCN	0.286	4	nearest	IT-STGCN	1.262	0.066
37	TGCN	0.286	4	nearest	STGCN	1.232	0.069

W_st

pd.merge(df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4 and mrate==0.3")

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.3	4	linear	IT-STGCN	1.153	0.036
1	DCRNN	0.3	4	linear	STGCN	1.263	0.053
2	DCRNN	0.3	4	nearest	IT-STGCN	1.154	0.038
3	DCRNN	0.3	4	nearest	STGCN	1.269	0.068
8	DyGrEncoder	0.3	4	linear	IT-STGCN	1.222	0.083
9	DyGrEncoder	0.3	4	linear	STGCN	1.276	0.058
10	DyGrEncoder	0.3	4	nearest	IT-STGCN	1.208	0.091
11	DyGrEncoder	0.3	4	nearest	STGCN	1.281	0.068
16	EvolveGCNH	0.3	4	linear	IT-STGCN	1.218	0.058
17	EvolveGCNH	0.3	4	linear	STGCN	1.237	0.051
18	EvolveGCNH	0.3	4	nearest	IT-STGCN	1.197	0.068
19	EvolveGCNH	0.3	4	nearest	STGCN	1.237	0.058
24	EvolveGCNO	0.3	4	linear	IT-STGCN	1.223	0.041
25	EvolveGCNO	0.3	4	linear	STGCN	1.263	0.048
26	EvolveGCNO	0.3	4	nearest	IT-STGCN	1.234	0.046
27	EvolveGCNO	0.3	4	nearest	STGCN	1.252	0.071
32	GCLSTM	0.3	4	linear	IT-STGCN	1.191	0.041
33	GCLSTM	0.3	4	linear	STGCN	1.264	0.041
34	GCLSTM	0.3	4	nearest	IT-STGCN	1.193	0.033
35	GCLSTM	0.3	4	nearest	STGCN	1.250	0.049
40	GConvGRU	0.3	4	linear	IT-STGCN	1.270	0.163
41	GConvGRU	0.3	4	linear	STGCN	1.556	0.264
42	GConvGRU	0.3	4	nearest	IT-STGCN	1.324	0.163
43	GConvGRU	0.3	4	nearest	STGCN	1.520	0.206
48	GConvLSTM	0.3	4	linear	IT-STGCN	1.340	0.166
49	GConvLSTM	0.3	4	linear	STGCN	1.392	0.109
50	GConvLSTM	0.3	4	nearest	IT-STGCN	1.368	0.158
51	GConvLSTM	0.3	4	nearest	STGCN	1.338	0.118
56	LRGCN	0.3	4	linear	IT-STGCN	1.299	0.147
57	LRGCN	0.3	4	linear	STGCN	1.325	0.086
58	LRGCN	0.3	4	nearest	IT-STGCN	1.260	0.117
59	LRGCN	0.3	4	nearest	STGCN	1.269	0.087
64	TGCN	0.3	4	linear	IT-STGCN	1.320	0.164
65	TGCN	0.3	4	linear	STGCN	1.287	0.126
66	TGCN	0.3	4	nearest	IT-STGCN	1.276	0.105
67	TGCN	0.3	4	nearest	STGCN	1.313	0.101

테이블 정리용

pd.merge(df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4 and mrate == 0.6 and inter_method=='linear'")

	model	mrate	lags	inter_method	method	mean	std
4	DCRNN	0.6	4	linear	IT-STGCN	1.241	0.079
5	DCRNN	0.6	4	linear	STGCN	1.506	0.065
12	DyGrEncoder	0.6	4	linear	IT-STGCN	1.287	0.095
13	DyGrEncoder	0.6	4	linear	STGCN	1.497	0.077
20	EvolveGCNH	0.6	4	linear	IT-STGCN	1.229	0.070
21	EvolveGCNH	0.6	4	linear	STGCN	1.278	0.066
28	EvolveGCNO	0.6	4	linear	IT-STGCN	1.269	0.092
29	EvolveGCNO	0.6	4	linear	STGCN	1.304	0.061
36	GCLSTM	0.6	4	linear	IT-STGCN	1.260	0.084
37	GCLSTM	0.6	4	linear	STGCN	1.340	0.059
44	GConvGRU	0.6	4	linear	IT-STGCN	1.434	0.222
45	GConvGRU	0.6	4	linear	STGCN	1.678	0.211
52	GConvLSTM	0.6	4	linear	IT-STGCN	1.312	0.162
53	GConvLSTM	0.6	4	linear	STGCN	1.498	0.083
60	LRGCN	0.6	4	linear	IT-STGCN	1.265	0.100
61	LRGCN	0.6	4	linear	STGCN	1.466	0.085
68	TGCN	0.6	4	linear	IT-STGCN	1.304	0.129
69	TGCN	0.6	4	linear	STGCN	1.299	0.076

pd.merge(df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("method !='STGCN' and lags==4 and mrate == 0.6 and inter_method=='linear'")

	model	mrate	lags	inter_method	method	mean	std
4	DCRNN	0.6	4	linear	IT-STGCN	1.241	0.079
12	DyGrEncoder	0.6	4	linear	IT-STGCN	1.287	0.095
20	EvolveGCNH	0.6	4	linear	IT-STGCN	1.229	0.070
28	EvolveGCNO	0.6	4	linear	IT-STGCN	1.269	0.092
36	GCLSTM	0.6	4	linear	IT-STGCN	1.260	0.084
44	GConvGRU	0.6	4	linear	IT-STGCN	1.434	0.222
52	GConvLSTM	0.6	4	linear	IT-STGCN	1.312	0.162
60	LRGCN	0.6	4	linear	IT-STGCN	1.265	0.100
68	TGCN	0.6	4	linear	IT-STGCN	1.304	0.129

pd.merge(df.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("method !='STGCN' and lags==4 and mrate == 0.6 and inter_method=='linear'")

	model	mrate	lags	inter_method	method	mean	std
8	DCRNN	0.6	4	linear	IT-STGCN	1.314	0.072
24	DyGrEncoder	0.6	4	linear	IT-STGCN	1.294	0.056
40	EvolveGCNH	0.6	4	linear	IT-STGCN	1.279	0.057
56	EvolveGCNO	0.6	4	linear	IT-STGCN	1.280	0.065
72	GCLSTM	0.6	4	linear	IT-STGCN	1.278	0.040
88	GConvGRU	0.6	4	linear	IT-STGCN	1.516	0.211
104	GConvLSTM	0.6	4	linear	IT-STGCN	1.255	0.049
116	GNAR	0.6	4	linear	GNAR	1.303	0.000
128	LRGCN	0.6	4	linear	IT-STGCN	1.286	0.053
144	TGCN	0.6	4	linear	IT-STGCN	1.257	0.048

pd.merge(df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4 and mrate == 0.6")

	model	mrate	lags	inter_method	method	mean	std
4	DCRNN	0.6	4	linear	IT-STGCN	1.241	0.079
5	DCRNN	0.6	4	linear	STGCN	1.506	0.065
6	DCRNN	0.6	4	nearest	IT-STGCN	1.208	0.079
7	DCRNN	0.6	4	nearest	STGCN	1.552	0.087
12	DyGrEncoder	0.6	4	linear	IT-STGCN	1.287	0.095
13	DyGrEncoder	0.6	4	linear	STGCN	1.497	0.077
14	DyGrEncoder	0.6	4	nearest	IT-STGCN	1.305	0.131
15	DyGrEncoder	0.6	4	nearest	STGCN	1.513	0.073
20	EvolveGCNH	0.6	4	linear	IT-STGCN	1.229	0.070
21	EvolveGCNH	0.6	4	linear	STGCN	1.278	0.066
22	EvolveGCNH	0.6	4	nearest	IT-STGCN	1.246	0.067
23	EvolveGCNH	0.6	4	nearest	STGCN	1.291	0.063
28	EvolveGCNO	0.6	4	linear	IT-STGCN	1.269	0.092
29	EvolveGCNO	0.6	4	linear	STGCN	1.304	0.061
30	EvolveGCNO	0.6	4	nearest	IT-STGCN	1.248	0.072
31	EvolveGCNO	0.6	4	nearest	STGCN	1.321	0.094
36	GCLSTM	0.6	4	linear	IT-STGCN	1.260	0.084
37	GCLSTM	0.6	4	linear	STGCN	1.340	0.059
38	GCLSTM	0.6	4	nearest	IT-STGCN	1.231	0.044
39	GCLSTM	0.6	4	nearest	STGCN	1.355	0.068
44	GConvGRU	0.6	4	linear	IT-STGCN	1.434	0.222
45	GConvGRU	0.6	4	linear	STGCN	1.678	0.211
46	GConvGRU	0.6	4	nearest	IT-STGCN	1.410	0.208
47	GConvGRU	0.6	4	nearest	STGCN	1.771	0.220
52	GConvLSTM	0.6	4	linear	IT-STGCN	1.312	0.162
53	GConvLSTM	0.6	4	linear	STGCN	1.498	0.083
54	GConvLSTM	0.6	4	nearest	IT-STGCN	1.313	0.205
55	GConvLSTM	0.6	4	nearest	STGCN	1.503	0.101
60	LRGCN	0.6	4	linear	IT-STGCN	1.265	0.100
61	LRGCN	0.6	4	linear	STGCN	1.466	0.085
62	LRGCN	0.6	4	nearest	IT-STGCN	1.331	0.120
63	LRGCN	0.6	4	nearest	STGCN	1.453	0.115
68	TGCN	0.6	4	linear	IT-STGCN	1.304	0.129
69	TGCN	0.6	4	linear	STGCN	1.299	0.076
70	TGCN	0.6	4	nearest	IT-STGCN	1.338	0.202
71	TGCN	0.6	4	nearest	STGCN	1.297	0.093

conclusion table

pd.merge(df2.query("dataset == 'pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4 and inter_method=='linear'")

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.286	4	linear	IT-STGCN	1.145	0.013
1	DCRNN	0.286	4	linear	STGCN	1.295	0.019
4	DyGrEncoder	0.286	4	linear	IT-STGCN	1.196	0.055
5	DyGrEncoder	0.286	4	linear	STGCN	1.224	0.037
8	EvolveGCNH	0.286	4	linear	IT-STGCN	1.188	0.042
9	EvolveGCNH	0.286	4	linear	STGCN	1.230	0.056
12	EvolveGCNO	0.286	4	linear	IT-STGCN	1.204	0.033
13	EvolveGCNO	0.286	4	linear	STGCN	1.210	0.058
16	GCLSTM	0.286	4	linear	IT-STGCN	1.182	0.045
17	GCLSTM	0.286	4	linear	STGCN	1.225	0.030
20	GConvGRU	0.286	4	linear	IT-STGCN	1.391	0.151
21	GConvGRU	0.286	4	linear	STGCN	1.420	0.110
24	GConvLSTM	0.286	4	linear	IT-STGCN	1.329	0.120
25	GConvLSTM	0.286	4	linear	STGCN	1.372	0.199
28	LRGCN	0.286	4	linear	IT-STGCN	1.201	0.081
29	LRGCN	0.286	4	linear	STGCN	1.227	0.070
32	TGCN	0.286	4	linear	IT-STGCN	1.243	0.110
33	TGCN	0.286	4	linear	STGCN	1.176	0.068

pd.merge(df2.query("dataset == 'pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df2.query("dataset == 'pedalme' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags','inter_method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("lags==4")

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.286	4	linear	IT-STGCN	1.145	0.013
1	DCRNN	0.286	4	linear	STGCN	1.295	0.019
2	DCRNN	0.286	4	nearest	IT-STGCN	1.143	0.011
3	DCRNN	0.286	4	nearest	STGCN	1.310	0.019
4	DyGrEncoder	0.286	4	linear	IT-STGCN	1.196	0.055
5	DyGrEncoder	0.286	4	linear	STGCN	1.224	0.037
6	DyGrEncoder	0.286	4	nearest	IT-STGCN	1.204	0.063
7	DyGrEncoder	0.286	4	nearest	STGCN	1.246	0.043
8	EvolveGCNH	0.286	4	linear	IT-STGCN	1.188	0.042
9	EvolveGCNH	0.286	4	linear	STGCN	1.230	0.056
10	EvolveGCNH	0.286	4	nearest	IT-STGCN	1.195	0.037
11	EvolveGCNH	0.286	4	nearest	STGCN	1.240	0.062
12	EvolveGCNO	0.286	4	linear	IT-STGCN	1.204	0.033
13	EvolveGCNO	0.286	4	linear	STGCN	1.210	0.058
14	EvolveGCNO	0.286	4	nearest	IT-STGCN	1.211	0.033
15	EvolveGCNO	0.286	4	nearest	STGCN	1.241	0.095
16	GCLSTM	0.286	4	linear	IT-STGCN	1.182	0.045
17	GCLSTM	0.286	4	linear	STGCN	1.225	0.030
18	GCLSTM	0.286	4	nearest	IT-STGCN	1.185	0.035
19	GCLSTM	0.286	4	nearest	STGCN	1.249	0.027
20	GConvGRU	0.286	4	linear	IT-STGCN	1.391	0.151
21	GConvGRU	0.286	4	linear	STGCN	1.420	0.110
22	GConvGRU	0.286	4	nearest	IT-STGCN	1.361	0.114
23	GConvGRU	0.286	4	nearest	STGCN	1.430	0.145
24	GConvLSTM	0.286	4	linear	IT-STGCN	1.329	0.120
25	GConvLSTM	0.286	4	linear	STGCN	1.372	0.199
26	GConvLSTM	0.286	4	nearest	IT-STGCN	1.310	0.151
27	GConvLSTM	0.286	4	nearest	STGCN	1.459	0.153
28	LRGCN	0.286	4	linear	IT-STGCN	1.201	0.081
29	LRGCN	0.286	4	linear	STGCN	1.227	0.070
30	LRGCN	0.286	4	nearest	IT-STGCN	1.197	0.106
31	LRGCN	0.286	4	nearest	STGCN	1.347	0.117
32	TGCN	0.286	4	linear	IT-STGCN	1.243	0.110
33	TGCN	0.286	4	linear	STGCN	1.176	0.068
34	TGCN	0.286	4	nearest	IT-STGCN	1.237	0.083
35	TGCN	0.286	4	nearest	STGCN	1.258	0.064

WikiMathsDatasetLoader (lags=8)

df.query("model=='GNAR' and dataset=='wikimath'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26545	wikimath	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.353637	228.561262	GNAR
26546	wikimath	GNAR	0.300000	rand	8	NaN	linear	NaN	1.353637	214.543727	GNAR
26547	wikimath	GNAR	0.800000	rand	8	NaN	linear	NaN	1.353637	204.452467	GNAR
26548	wikimath	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.353637	129.201734	GNAR
26549	wikimath	GNAR	0.300000	rand	8	NaN	linear	NaN	1.353637	193.721538	GNAR
26550	wikimath	GNAR	0.800000	rand	8	NaN	linear	NaN	1.353637	188.391585	GNAR
26551	wikimath	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.353637	115.926373	GNAR
26552	wikimath	GNAR	0.300000	rand	8	NaN	linear	NaN	1.353637	223.202336	GNAR
26553	wikimath	GNAR	0.800000	rand	8	NaN	linear	NaN	1.353637	126.158856	GNAR
26578	wikimath	GNAR	0.119837	block	8	NaN	linear	NaN	1.353637	114.421239	GNAR
26579	wikimath	GNAR	0.119837	block	8	NaN	linear	NaN	1.353637	214.541294	GNAR
26580	wikimath	GNAR	0.119837	block	8	NaN	linear	NaN	1.353637	269.699585	GNAR

Baseline

pd.merge(df.query("dataset=='wikimath' and mrate==0").groupby(['model','lags','nof_filters','method'])['mse'].mean().reset_index(),
         df.query("dataset=='wikimath' and mrate==0").groupby(['model','lags','nof_filters','method'])['mse'].std().reset_index(),
         on=['model','lags','nof_filters','method']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	lags	nof_filters	method	mean	std
0	DCRNN	8	12.0	IT-STGCN	0.582	0.006
1	DCRNN	8	12.0	STGCN	0.580	0.006
2	DyGrEncoder	8	12.0	IT-STGCN	0.563	0.030
3	DyGrEncoder	8	12.0	STGCN	0.560	0.029
4	EvolveGCNH	8	12.0	IT-STGCN	0.784	0.027
5	EvolveGCNH	8	12.0	STGCN	0.771	0.028
6	EvolveGCNO	8	12.0	IT-STGCN	0.735	0.023
7	EvolveGCNO	8	12.0	STGCN	0.734	0.025
8	GCLSTM	8	64.0	IT-STGCN	0.643	0.024
9	GCLSTM	8	64.0	STGCN	0.645	0.018
10	GConvGRU	8	12.0	IT-STGCN	0.529	0.003
11	GConvGRU	8	12.0	STGCN	0.528	0.003
12	GConvLSTM	8	64.0	IT-STGCN	0.626	0.015
13	GConvLSTM	8	64.0	STGCN	0.640	0.031
14	LRGCN	8	32.0	IT-STGCN	0.610	0.017
15	LRGCN	8	32.0	STGCN	0.608	0.014
16	TGCN	8	12.0	IT-STGCN	0.730	0.030
17	TGCN	8	12.0	STGCN	0.732	0.035

Random

fig, ax = plt.subplots(3, 3,figsize=(40,20))

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==8 and epoch==50 and model=='GConvGRU'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,0],grid=False,widths=0.5)
ax[0,0].set_title('GConvGRU')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==64 and lags==8 and epoch==50 and model=='GConvLSTM'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,1],grid=False,widths=0.5)
ax[0,1].set_title('GConvLSTM')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==64 and lags==8 and epoch==50 and model=='GCLSTM'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[0,2],grid=False,widths=0.5)
ax[0,2].set_title('GCLSTM')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==32 and lags==8 and epoch==50 and model=='LRGCN'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,0],grid=False,widths=0.5)
ax[1,0].set_title('LRGCN')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==8 and epoch==50 and model=='DyGrEncoder'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,1],grid=False,widths=0.5)
ax[1,1].set_title('DyGrEncoder')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and lags==8 and epoch==50 and model=='EvolveGCNH'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[1,2],grid=False,widths=0.5)
ax[1,2].set_title('EvolveGCNH')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and lags==8 and epoch==50 and model=='EvolveGCNO'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,0],grid=False,widths=0.5)
ax[2,0].set_title('EvolveGCNO')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==8 and epoch==50 and model=='TGCN'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,1],grid=False,widths=0.5)
ax[2,1].set_title('TGCN')

df.query("dataset=='wikimath' and mtype=='rand' and inter_method == 'linear' and nof_filters==12 and lags==8 and epoch==50 and model=='DCRNN'").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax[2,2],grid=False,widths=0.5)
ax[2,2].set_title('DCRNN')


for ax in ax.flat:
    ax.set_yticklabels([])
    ax.axvline(x=2.5, color='black', linestyle='-')
    ax.axvline(x=4.5, color='black', linestyle='-')
    ax.axvline(x=6.5, color='black', linestyle='-')
    ax.axvline(x=8.5, color='black', linestyle='-')
    # ax.set_xticklabels(['IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN','IT-TGNN','TGNN'])
    ax.set_xlabel('')
    ax.set_ylabel('')
    
fig.suptitle('',fontsize=40)

Text(0.5, 0.98, '')

pd.merge(df.query("dataset=='wikimath' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='wikimath' and mtype=='rand'").groupby(['model','mrate','lags','nof_filters','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','nof_filters','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate == 0.3")

	model	mrate	lags	nof_filters	inter_method	method	mean	std
0	DCRNN	0.3	8	12.0	linear	IT-STGCN	0.588	0.007
1	DCRNN	0.3	8	12.0	linear	STGCN	0.603	0.010
8	DyGrEncoder	0.3	8	12.0	linear	IT-STGCN	0.578	0.031
9	DyGrEncoder	0.3	8	12.0	linear	STGCN	0.562	0.016
16	EvolveGCNH	0.3	8	12.0	linear	IT-STGCN	0.775	0.021
17	EvolveGCNH	0.3	8	12.0	linear	STGCN	0.787	0.024
24	EvolveGCNO	0.3	8	12.0	linear	IT-STGCN	0.738	0.018
25	EvolveGCNO	0.3	8	12.0	linear	STGCN	0.743	0.024
32	GCLSTM	0.3	8	64.0	linear	IT-STGCN	0.628	0.020
33	GCLSTM	0.3	8	64.0	linear	STGCN	0.674	0.020
40	GConvGRU	0.3	8	12.0	linear	IT-STGCN	0.518	0.002
41	GConvGRU	0.3	8	12.0	linear	STGCN	0.570	0.006
48	GConvLSTM	0.3	8	64.0	linear	IT-STGCN	0.631	0.019
49	GConvLSTM	0.3	8	64.0	linear	STGCN	0.764	0.057
56	LRGCN	0.3	8	32.0	linear	IT-STGCN	0.619	0.019
57	LRGCN	0.3	8	32.0	linear	STGCN	0.689	0.032
64	TGCN	0.3	8	12.0	linear	IT-STGCN	0.739	0.040
65	TGCN	0.3	8	12.0	linear	STGCN	0.734	0.027

pd.merge(df.query("dataset=='wikimath' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='wikimath' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("mrate != 0.3")

	model	mrate	lags	inter_method	method	mean	std
2	DCRNN	0.5	8	linear	IT-STGCN	0.590	0.006
3	DCRNN	0.5	8	linear	STGCN	0.652	0.015
4	DCRNN	0.6	8	linear	IT-STGCN	0.592	0.005
5	DCRNN	0.6	8	linear	STGCN	0.688	0.011
6	DCRNN	0.8	8	linear	IT-STGCN	0.672	0.007
7	DCRNN	0.8	8	linear	STGCN	0.846	0.031
10	DyGrEncoder	0.5	8	linear	IT-STGCN	0.565	0.024
11	DyGrEncoder	0.5	8	linear	STGCN	0.614	0.024
12	DyGrEncoder	0.6	8	linear	IT-STGCN	0.566	0.021
13	DyGrEncoder	0.6	8	linear	STGCN	0.644	0.032
14	DyGrEncoder	0.8	8	linear	IT-STGCN	0.606	0.017
15	DyGrEncoder	0.8	8	linear	STGCN	0.770	0.045
18	EvolveGCNH	0.5	8	linear	IT-STGCN	0.791	0.029
19	EvolveGCNH	0.5	8	linear	STGCN	0.804	0.026
20	EvolveGCNH	0.6	8	linear	IT-STGCN	0.795	0.025
21	EvolveGCNH	0.6	8	linear	STGCN	0.843	0.042
22	EvolveGCNH	0.8	8	linear	IT-STGCN	0.877	0.045
23	EvolveGCNH	0.8	8	linear	STGCN	0.915	0.063
26	EvolveGCNO	0.5	8	linear	IT-STGCN	0.744	0.021
27	EvolveGCNO	0.5	8	linear	STGCN	0.759	0.021
28	EvolveGCNO	0.6	8	linear	IT-STGCN	0.745	0.019
29	EvolveGCNO	0.6	8	linear	STGCN	0.775	0.026
30	EvolveGCNO	0.8	8	linear	IT-STGCN	0.780	0.027
31	EvolveGCNO	0.8	8	linear	STGCN	0.863	0.038
34	GCLSTM	0.5	8	linear	IT-STGCN	0.626	0.010
35	GCLSTM	0.5	8	linear	STGCN	0.825	0.045
36	GCLSTM	0.6	8	linear	IT-STGCN	0.650	0.021
37	GCLSTM	0.6	8	linear	STGCN	0.942	0.078
38	GCLSTM	0.8	8	linear	IT-STGCN	0.815	0.058
39	GCLSTM	0.8	8	linear	STGCN	1.407	0.117
42	GConvGRU	0.5	8	linear	IT-STGCN	0.524	0.003
43	GConvGRU	0.5	8	linear	STGCN	0.658	0.010
44	GConvGRU	0.6	8	linear	IT-STGCN	0.539	0.004
45	GConvGRU	0.6	8	linear	STGCN	0.731	0.015
46	GConvGRU	0.8	8	linear	IT-STGCN	0.687	0.021
47	GConvGRU	0.8	8	linear	STGCN	0.932	0.043
50	GConvLSTM	0.5	8	linear	IT-STGCN	0.669	0.034
51	GConvLSTM	0.5	8	linear	STGCN	0.965	0.077
52	GConvLSTM	0.6	8	linear	IT-STGCN	0.731	0.048
53	GConvLSTM	0.6	8	linear	STGCN	1.104	0.078
54	GConvLSTM	0.8	8	linear	IT-STGCN	0.920	0.069
55	GConvLSTM	0.8	8	linear	STGCN	1.423	0.121
57	GNAR	0.8	8	linear	GNAR	1.354	0.000
60	LRGCN	0.5	8	linear	IT-STGCN	0.621	0.015
61	LRGCN	0.5	8	linear	STGCN	0.838	0.054
62	LRGCN	0.6	8	linear	IT-STGCN	0.648	0.024
63	LRGCN	0.6	8	linear	STGCN	0.917	0.063
64	LRGCN	0.8	8	linear	IT-STGCN	0.769	0.045
65	LRGCN	0.8	8	linear	STGCN	1.105	0.099
68	TGCN	0.5	8	linear	IT-STGCN	0.757	0.046
69	TGCN	0.5	8	linear	STGCN	0.757	0.034
70	TGCN	0.6	8	linear	IT-STGCN	0.742	0.030
71	TGCN	0.6	8	linear	STGCN	0.774	0.025
72	TGCN	0.8	8	linear	IT-STGCN	0.771	0.020
73	TGCN	0.8	8	linear	STGCN	0.827	0.030

Block

pd.merge(df.query("dataset=='wikimath' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='wikimath' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.120	8	linear	IT-STGCN	0.583	0.006
1	DCRNN	0.120	8	linear	STGCN	0.578	0.005
2	DyGrEncoder	0.120	8	linear	IT-STGCN	0.563	0.025
3	DyGrEncoder	0.120	8	linear	STGCN	0.546	0.016
4	EvolveGCNH	0.120	8	linear	IT-STGCN	0.776	0.028
5	EvolveGCNH	0.120	8	linear	STGCN	0.773	0.021
6	EvolveGCNO	0.120	8	linear	IT-STGCN	0.732	0.025
7	EvolveGCNO	0.120	8	linear	STGCN	0.735	0.022
8	GCLSTM	0.120	8	linear	IT-STGCN	0.640	0.019
9	GCLSTM	0.120	8	linear	STGCN	0.638	0.013
10	GConvGRU	0.004	8	linear	IT-STGCN	0.529	0.003
11	GConvGRU	0.004	8	linear	STGCN	0.528	0.003
12	GConvGRU	0.096	8	linear	IT-STGCN	0.529	0.004
13	GConvGRU	0.096	8	linear	STGCN	0.544	0.011
14	GConvGRU	0.120	8	linear	IT-STGCN	0.523	0.002
15	GConvGRU	0.120	8	linear	STGCN	0.531	0.002
16	GConvLSTM	0.120	8	linear	IT-STGCN	0.627	0.014
17	GConvLSTM	0.120	8	linear	STGCN	0.660	0.034
18	GNAR	0.120	8	linear	GNAR	1.354	0.000
19	LRGCN	0.120	8	linear	IT-STGCN	0.608	0.012
20	LRGCN	0.120	8	linear	STGCN	0.624	0.024
21	TGCN	0.120	8	linear	IT-STGCN	0.748	0.046
22	TGCN	0.120	8	linear	STGCN	0.741	0.046

missing values on the same nodes

pd.merge(df2.query("dataset=='wikimath'").groupby(['model','mrate','lags','method'])['mse'].mean().reset_index(),
        df2.query("dataset=='wikimath'").groupby(['model','mrate','lags','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	lags	method	mean	std
0	DCRNN	0.512	8	IT-STGCN	0.592	0.005
1	DCRNN	0.512	8	STGCN	0.665	0.015
2	DyGrEncoder	0.512	8	IT-STGCN	0.561	0.031
3	DyGrEncoder	0.512	8	STGCN	0.626	0.027
4	EvolveGCNH	0.512	8	IT-STGCN	0.794	0.031
5	EvolveGCNH	0.512	8	STGCN	0.818	0.031
6	EvolveGCNO	0.512	8	IT-STGCN	0.745	0.017
7	EvolveGCNO	0.512	8	STGCN	0.753	0.026
8	GCLSTM	0.512	8	IT-STGCN	0.617	0.011
9	GCLSTM	0.512	8	STGCN	0.823	0.048
10	GConvGRU	0.512	8	IT-STGCN	0.533	0.003
11	GConvGRU	0.512	8	STGCN	0.726	0.015
12	GConvLSTM	0.512	8	IT-STGCN	0.653	0.033
13	GConvLSTM	0.512	8	STGCN	0.963	0.098
14	GNAR	0.512	8	GNAR	1.354	0.000
15	LRGCN	0.512	8	IT-STGCN	0.624	0.019
16	LRGCN	0.512	8	STGCN	0.810	0.064
17	TGCN	0.512	8	IT-STGCN	0.750	0.039
18	TGCN	0.512	8	STGCN	0.782	0.030

Windmillsmall(lags=8)

df.query("model=='GNAR' and dataset=='windmillsmall'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26584	windmillsmall	GNAR	0.081291	block	8	NaN	linear	NaN	1.64923	3.428970	GNAR
26585	windmillsmall	GNAR	0.081291	block	8	NaN	linear	NaN	1.64923	3.492012	GNAR
26586	windmillsmall	GNAR	0.081291	block	8	NaN	linear	NaN	1.64923	2.916400	GNAR
26587	windmillsmall	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.64923	2.513212	GNAR
26588	windmillsmall	GNAR	0.700000	rand	8	NaN	nearest	NaN	1.64923	3.334994	GNAR
26589	windmillsmall	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.64923	2.116200	GNAR
26590	windmillsmall	GNAR	0.700000	rand	8	NaN	nearest	NaN	1.64923	2.391699	GNAR
26591	windmillsmall	GNAR	0.000000	NaN	8	NaN	NaN	NaN	1.64923	3.406726	GNAR
26592	windmillsmall	GNAR	0.700000	rand	8	NaN	nearest	NaN	1.64923	2.586957	GNAR

Baseline

pd.merge(df.query("dataset=='windmillsmall' and mrate==0").groupby(['model','lags'])['mse'].mean().reset_index(),
         df.query("dataset=='windmillsmall' and mrate==0").groupby(['model','lags'])['mse'].std().reset_index(),
         on=['model','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	lags	mean	std
0	DCRNN	8	0.988	0.003
1	DyGrEncoder	8	0.988	0.008
2	EvolveGCNH	8	0.986	0.002
3	EvolveGCNO	8	0.983	0.001
4	GCLSTM	8	0.992	0.010
5	GConvGRU	8	1.003	0.004
6	GConvLSTM	8	1.019	0.045
7	GNAR	8	1.649	0.000
8	LRGCN	8	0.987	0.006
9	TGCN	8	0.991	0.010

Random

fig, ((ax1)) = plt.subplots(nrows=1, ncols=1, figsize=(50, 6), sharey=True)
df.query("dataset=='windmillsmall' and mtype=='rand' and inter_method == 'linear' and lags==8 and epoch==50").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax1,grid=False,widths=0.5)

pd.merge(df.query("dataset=='windmillsmall' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='windmillsmall' and mtype=='rand'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.7	8	linear	IT-STGCN	1.117	0.034
1	DCRNN	0.7	8	linear	STGCN	1.348	0.057
2	DyGrEncoder	0.7	8	linear	IT-STGCN	1.127	0.009
3	DyGrEncoder	0.7	8	linear	STGCN	1.709	0.074
4	EvolveGCNH	0.7	8	linear	IT-STGCN	1.129	0.035
5	EvolveGCNH	0.7	8	linear	STGCN	1.330	0.137
6	EvolveGCNO	0.7	8	linear	IT-STGCN	1.149	0.026
7	EvolveGCNO	0.7	8	linear	STGCN	1.495	0.137
8	GCLSTM	0.7	8	linear	IT-STGCN	1.116	0.021
9	GCLSTM	0.7	8	linear	STGCN	1.573	0.105
10	GConvGRU	0.7	8	linear	IT-STGCN	1.194	0.042
11	GConvGRU	0.7	8	linear	STGCN	1.662	0.073
12	GConvLSTM	0.7	8	linear	IT-STGCN	1.142	0.021
13	GConvLSTM	0.7	8	linear	STGCN	1.599	0.057
14	GNAR	0.7	8	nearest	GNAR	1.649	0.000
15	LRGCN	0.7	8	linear	IT-STGCN	1.110	0.012
16	LRGCN	0.7	8	linear	STGCN	1.492	0.087
17	TGCN	0.7	8	linear	IT-STGCN	1.071	0.010
18	TGCN	0.7	8	linear	STGCN	1.305	0.039

Block

pd.merge(df.query("dataset=='windmillsmall' and mtype=='block'").groupby(['model','mrate','nof_filters','lags','method'])['mse'].mean().reset_index(),
         df.query("dataset=='windmillsmall' and mtype=='block'").groupby(['model','mrate','nof_filters','lags','method'])['mse'].std().reset_index(),
         on=['model','method','nof_filters','mrate','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	nof_filters	lags	method	mean	std
0	DCRNN	0.081	4.0	8	IT-STGCN	0.983	0.002
1	DCRNN	0.081	4.0	8	STGCN	0.994	0.005
2	DyGrEncoder	0.081	12.0	8	IT-STGCN	0.985	0.005
3	DyGrEncoder	0.081	12.0	8	STGCN	0.985	0.003
4	EvolveGCNH	0.081	12.0	8	IT-STGCN	0.986	0.003
5	EvolveGCNH	0.081	12.0	8	STGCN	0.993	0.003
6	EvolveGCNO	0.081	12.0	8	IT-STGCN	0.983	0.002
7	EvolveGCNO	0.081	12.0	8	STGCN	0.990	0.002
8	GCLSTM	0.081	16.0	8	IT-STGCN	0.985	0.003
9	GCLSTM	0.081	16.0	8	STGCN	0.985	0.002
10	GConvGRU	0.081	12.0	8	IT-STGCN	1.007	0.005
11	GConvGRU	0.081	12.0	8	STGCN	1.008	0.006
12	GConvLSTM	0.081	16.0	8	IT-STGCN	0.997	0.022
13	GConvLSTM	0.081	16.0	8	STGCN	0.989	0.009
14	LRGCN	0.081	12.0	8	IT-STGCN	0.985	0.002
15	LRGCN	0.081	12.0	8	STGCN	0.985	0.003
16	TGCN	0.081	12.0	8	IT-STGCN	0.992	0.015
17	TGCN	0.081	12.0	8	STGCN	0.999	0.013

Montevideobus (lags=4)

df.query("model=='GNAR' and dataset=='monte'")

	dataset	method	mrate	mtype	lags	nof_filters	inter_method	epoch	mse	calculation_time	model
26554	monte	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.061937	6.576898	GNAR
26555	monte	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.061937	4.220497	GNAR
26556	monte	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.061937	4.429234	GNAR
26557	monte	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.061937	4.886660	GNAR
26558	monte	GNAR	0.000000	NaN	4	NaN	NaN	NaN	1.061937	4.879432	GNAR
26559	monte	GNAR	0.800000	rand	4	NaN	nearest	NaN	1.061937	4.814065	GNAR
26581	monte	GNAR	0.149142	block	4	NaN	nearest	NaN	1.061937	4.948565	GNAR
26582	monte	GNAR	0.149142	block	4	NaN	nearest	NaN	1.061937	2.942073	GNAR
26583	monte	GNAR	0.149142	block	4	NaN	nearest	NaN	1.061937	5.179985	GNAR

Baseline

pd.merge(df.query("dataset=='monte' and mrate==0").groupby(['model','lags'])['mse'].mean().reset_index(),
         df.query("dataset=='monte' and mrate==0").groupby(['model','lags'])['mse'].std().reset_index(),
         on=['model','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	lags	mean	std
0	DCRNN	4	0.936	0.002
1	DyGrEncoder	4	0.995	0.034
2	EvolveGCNH	4	1.182	0.192
3	EvolveGCNO	4	1.157	0.182
4	GCLSTM	4	0.970	0.011
5	GConvGRU	4	0.931	0.002
6	GConvLSTM	4	0.960	0.011
7	GNAR	4	1.062	0.000
8	LRGCN	4	0.980	0.024
9	TGCN	4	0.983	0.006

Random

fig, ((ax1)) = plt.subplots(nrows=1, ncols=1, figsize=(50, 6), sharey=True)
df.query("dataset=='monte' and mtype=='rand' and inter_method == 'nearest' and lags==4 and epoch==50").\
iloc[:,[1,2,8,10]].boxplot(by=['model','mrate','method'],ax=ax1,grid=False,widths=0.5)

pd.merge(df.query("dataset=='monte' and mtype=='rand'").groupby(['model','mrate','nof_filters','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='monte' and mtype=='rand'").groupby(['model','mrate','nof_filters','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','mrate','nof_filters','inter_method','method','mrate','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	nof_filters	lags	inter_method	method	mean	std
0	DCRNN	0.3	12.0	4	nearest	IT-STGCN	0.938	0.001
1	DCRNN	0.3	12.0	4	nearest	STGCN	1.008	0.006
2	DCRNN	0.5	12.0	4	nearest	IT-STGCN	0.948	0.002
3	DCRNN	0.5	12.0	4	nearest	STGCN	1.119	0.025
4	DCRNN	0.7	12.0	4	nearest	IT-STGCN	1.031	0.018
...	...	...	...	...	...	...	...	...
67	TGCN	0.5	8.0	4	nearest	STGCN	1.100	0.036
68	TGCN	0.7	8.0	4	nearest	IT-STGCN	1.030	0.010
69	TGCN	0.7	8.0	4	nearest	STGCN	1.183	0.057
70	TGCN	0.8	8.0	4	nearest	IT-STGCN	1.073	0.024
71	TGCN	0.8	8.0	4	nearest	STGCN	1.218	0.086

72 rows × 8 columns

Block

pd.merge(df.query("dataset=='monte' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='monte' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3)

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.149	4	nearest	IT-STGCN	0.940	0.001
1	DCRNN	0.149	4	nearest	STGCN	0.956	0.003
2	DyGrEncoder	0.149	4	nearest	IT-STGCN	1.005	0.046
3	DyGrEncoder	0.149	4	nearest	STGCN	1.030	0.044
4	EvolveGCNH	0.149	4	nearest	IT-STGCN	1.392	0.110
5	EvolveGCNH	0.149	4	nearest	STGCN	1.612	0.216
6	EvolveGCNO	0.149	4	nearest	IT-STGCN	1.345	0.110
7	EvolveGCNO	0.149	4	nearest	STGCN	1.766	0.123
8	GCLSTM	0.149	4	nearest	IT-STGCN	0.959	0.008
9	GCLSTM	0.149	4	nearest	STGCN	0.956	0.005
10	GConvGRU	0.149	4	cubic	IT-STGCN	1.023	0.021
11	GConvGRU	0.149	4	cubic	STGCN	1.028	0.031
12	GConvGRU	0.149	4	linear	IT-STGCN	0.930	0.002
13	GConvGRU	0.149	4	linear	STGCN	0.935	0.005
14	GConvGRU	0.149	4	nearest	IT-STGCN	0.932	0.002
15	GConvGRU	0.149	4	nearest	STGCN	0.935	0.004
16	GConvLSTM	0.149	4	nearest	IT-STGCN	0.949	0.008
17	GConvLSTM	0.149	4	nearest	STGCN	0.950	0.005
18	GNAR	0.149	4	nearest	GNAR	1.062	0.000
19	LRGCN	0.149	4	nearest	IT-STGCN	0.978	0.024
20	LRGCN	0.149	4	nearest	STGCN	0.977	0.020
21	TGCN	0.149	4	nearest	IT-STGCN	0.984	0.007
22	TGCN	0.149	4	nearest	STGCN	0.985	0.005

pd.merge(df.query("dataset=='monte' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].mean().reset_index(),
         df.query("dataset=='monte' and mtype=='block'").groupby(['model','mrate','lags','inter_method','method'])['mse'].std().reset_index(),
         on=['model','method','mrate','inter_method','lags']).rename(columns={'mse_x':'mean','mse_y':'std'}).round(3).query("inter_method=='nearest'")

	model	mrate	lags	inter_method	method	mean	std
0	DCRNN	0.149	4	nearest	IT-STGCN	0.940	0.001
1	DCRNN	0.149	4	nearest	STGCN	0.956	0.003
2	DyGrEncoder	0.149	4	nearest	IT-STGCN	1.005	0.046
3	DyGrEncoder	0.149	4	nearest	STGCN	1.030	0.044
4	EvolveGCNH	0.149	4	nearest	IT-STGCN	1.392	0.110
5	EvolveGCNH	0.149	4	nearest	STGCN	1.612	0.216
6	EvolveGCNO	0.149	4	nearest	IT-STGCN	1.345	0.110
7	EvolveGCNO	0.149	4	nearest	STGCN	1.766	0.123
8	GCLSTM	0.149	4	nearest	IT-STGCN	0.959	0.008
9	GCLSTM	0.149	4	nearest	STGCN	0.956	0.005
14	GConvGRU	0.149	4	nearest	IT-STGCN	0.932	0.002
15	GConvGRU	0.149	4	nearest	STGCN	0.935	0.004
16	GConvLSTM	0.149	4	nearest	IT-STGCN	0.949	0.008
17	GConvLSTM	0.149	4	nearest	STGCN	0.950	0.005
18	GNAR	0.149	4	nearest	GNAR	1.062	0.000
19	LRGCN	0.149	4	nearest	IT-STGCN	0.978	0.024
20	LRGCN	0.149	4	nearest	STGCN	0.977	0.020
21	TGCN	0.149	4	nearest	IT-STGCN	0.984	0.007
22	TGCN	0.149	4	nearest	STGCN	0.985	0.005